ExamSprintLab

200-901 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

The 200-901 exam lasts 120 minutes and costs 300 USD at the time of writing, according to Cisco’s exam page. Cisco does not publish the number of questions or a passing score on the pages used for this guide; the exam is graded pass or fail. It is offered in English and Japanese, and the CCNA Automation certification it earns is valid for three years.


Exam title	Automating Networks Using Cisco Platforms (200-901 CCNAAUTO)
Certification	CCNA Automation (previously DevNet Associate)
Level	Associate
Duration	120 minutes
Number of questions	Not published
Passing score	Not published (pass/fail)
Languages	English, Japanese
Price	300 USD at time of writing; Cisco Learning Credits accepted. Local taxes and currency may differ at booking
Prerequisites	None required
Validity	3 years
Recertification	Pass the current CCNA Automation exam or a higher-level Cisco exam, or earn 30 continuing education credits
Exam topics document	v1.1

Name and version

The code 200-901 has not changed. What changed is the name of the certification: the DevNet Associate became CCNA Automation in early 2026, and the exam code carries the label CCNAAUTO instead of DEVASC. Cisco’s learning blog states that DevNet became Cisco Automation, with CCNA, CCNP and CCIE naming from February 3, 2026, and describes the associate exam as “same test, new names”.

AB-100 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

AB-100 requires a scaled score of 700 out of 1000 to pass and is offered in 13 languages. Microsoft’s exam page does not publish the duration, the number of questions or a fixed price, and says only that the price depends on the country or region where you sit the exam. Below is everything the official pages state, and what it means for your revision.


Exam title	Agentic AI Business Solutions Architect
Certification	Microsoft Certified: Agentic AI Business Solutions Architect Expert
Level	Expert
Passing score	700 on a scale of 1–1000
Duration	Not published on the exam page
Number of questions	Not published by Microsoft
Price	Set per country or region; shown at booking
Languages	English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian, Spanish
Retirement date	None announced
Renewal	Annually, free, on Microsoft Learn
Current outline	Skills measured as of October 14, 2026

Check the booking screen for the duration and any extra-time options before you plan your day. Some third-party sites quote a time limit and question count; treat those as unverified until they appear on Microsoft’s own pages.

AB-730 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

AB-730 gives you 45 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Below is everything else it publishes, and what it means for your preparation. All of it comes from Microsoft’s certification page and study guide.


Exam title	AI Business Professional
Certification	Microsoft Certified: AI Business Professional
Level	Beginner
Duration	45 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian, Spanish
Delivery	Proctored, scheduled through Pearson VUE
Price	Set per country or region; shown at booking
Renewal	Not stated on the exam page; check your Microsoft Learn profile
Practice assessment	Available free on Microsoft Learn

Which outline this guide follows

Microsoft states that the English version of the exam is updated on October 20, 2026. This section is written to the newer outline, which is the one the study guide labels “skills measured as of October 20, 2026”. If you sit the English exam before that date, you are tested on the previous outline; if you sit it after, on the new one. If Microsoft revises the outline again, the study guide on Microsoft Learn is the authority and this section will be updated to match; see also exam changes.

AB-731 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

AB-731 gives you 45 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country or region at the booking step. Microsoft’s exam page does not publish the number of questions. The exam is proctored and is offered in 13 languages. Below is everything Microsoft publishes, and what it means for your preparation.


Exam title	AI Transformation Leader
Certification	Microsoft Certified: AI Transformation Leader
Level	Beginner
Role	Business leader
Duration	45 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian, Spanish
Delivery	Proctored; scheduled through Pearson VUE
Price	Set per country or region; shown at booking
Status	Generally available (no beta label on the exam page)
Outline last updated	July 22, 2026
Renewal	Not stated on the exam page; see the note below

The three skill areas

Skill area	Weight
Identify the business value of generative AI solutions	35–40%
Identify benefits, capabilities, and opportunities for Microsoft’s AI apps and services	35–40%
Identify an implementation and adoption strategy for Microsoft’s AI apps and services	20–25%

Every objective begins with “identify”, “describe”, “understand”, “explain” or “establish”. That sets the depth: you choose and justify, you do not build. The weights work out as follows for revision:

AZ-305 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

AZ-305 requires a scaled score of 700 out of 1000 to pass, is offered in ten languages, and is priced per country at the booking step. The exam page on Microsoft Learn does not state the duration or the number of questions, so check the exact time allowed in your appointment details when you book. This page lists what Microsoft publishes and what it means for your preparation.


Exam title	Designing Microsoft Azure Infrastructure Solutions
Certification	Microsoft Certified: Azure Solutions Architect Expert
Level	Expert
Duration	Not published on the exam page; confirm when booking
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Korean, German, French, Spanish, Portuguese (Brazil), Chinese (Traditional), Italian
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Prerequisite	Microsoft Certified: Azure Administrator Associate (AZ-104) to earn the certification
Renewal	Free online assessment on Microsoft Learn, annually
Current outline	Skills measured as of April 17, 2026

The four domains

Domain	Weight
Design identity, governance, and monitoring solutions	25–30%
Design data storage solutions	20–25%
Design business continuity solutions	15–20%
Design infrastructure solutions	30–35%

Infrastructure and identity together can make up well over half of the exam. Infrastructure is also the most crowded domain: compute, application architecture, migrations and networking each contain several objectives. Business continuity is the smallest domain, but its questions share vocabulary (RPO, RTO, zones, regions) with storage and infrastructure, so it is worth more than its weight suggests.

AZ-400 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

AZ-400 gives you 100 minutes, or 120 minutes if your exam includes labs, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish a question count for this exam. Below is everything it does publish, and what it means for your preparation.


Exam title	Designing and Implementing Microsoft DevOps Solutions
Certification	Microsoft Certified: DevOps Engineer Expert
Level	Expert
Duration	100 minutes; 120 minutes if the exam may contain labs (the time is shown when you register)
Passing score	700 on a scale of 1-1000
Number of questions	Not published for this exam. Microsoft says most exams have 40-60 questions, and the number can vary
Languages	English, Japanese, Chinese (Simplified), Korean, German, French, Spanish, Portuguese (Brazil), Chinese (Traditional), Italian
Delivery	Proctored, online or at a test centre
Price	Set per country or region; shown at booking
Renewal	Annual, free, through an online assessment on Microsoft Learn
Retirement	No retirement date published
Current outline	Skills measured as of July 27, 2026

The five domains

Domain	Weight
Design and implement processes and communications	10-15%
Design and implement a source control strategy	10-15%
Design and implement build and release pipelines	50-55%
Develop a security and compliance plan	10-15%
Implement an instrumentation strategy	5-10%

The pipelines domain dominates. It holds six sub-areas: package management, testing strategy, pipeline design, deployments, infrastructure as code and pipeline maintenance. Because that single domain is more than half of the exam, a candidate who knows every other domain perfectly but is weak on YAML, triggers, templates and deployment strategies will struggle to reach 700.

CCNA 200-301 exam format: score and cost

Fri, 09 Oct 2026 00:00:00 +0000

The CCNA 200-301 exam lasts 120 minutes and costs 300 USD at the time of writing, according to Cisco’s exam page. Cisco does not publish the number of questions or the passing score for this exam; it is graded pass or fail. The exam is offered in English and Japanese, and a passed CCNA is valid for three years.


Exam title	Implementing and Administering Cisco Solutions (200-301 CCNA)
Certification	Cisco Certified Network Associate (CCNA)
Level	Associate
Duration	120 minutes
Number of questions	Not published
Passing score	Not published (pass/fail)
Languages	English, Japanese on Cisco’s page for v1.1; Cisco has not published the v2.0 language list
Price	300 USD at time of writing; Cisco Learning Credits accepted. Local taxes and currency may differ at booking
Prerequisites	None required
Validity	3 years
Recertification	Pass the current CCNA exam or a higher-level Cisco exam, or earn 30 continuing education credits

Two versions, one exam code

The code 200-301 has covered several blueprint revisions. Two matter right now:

DP-600 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

DP-600 gives you 100 minutes, requires a scaled score of 700 on a scale of 1 to 1000, and is priced by the country or region where you take it. Microsoft does not publish the number of questions. The English exam changes on October 19, 2026, so check which outline applies to your booking date. Everything else Microsoft publishes is below.


Exam title	Implementing Analytics Solutions Using Microsoft Fabric
Certification	Microsoft Certified: Fabric Analytics Engineer Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), German, French, Spanish, Portuguese (Brazil)
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Current outline	Skills measured as of October 19, 2026 (English)

The October 19, 2026 change

The English version of the exam is updated on October 19, 2026, and Microsoft’s study guide publishes the skills measured as of that date. Localised versions follow later: the study guide says they are updated roughly eight weeks after English, though that schedule is not guaranteed. The previous version of the outline had four skill areas; this guide does not reproduce it, because the official page now shows only the new one.

DP-700 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

DP-700 gives you 100 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Implementing Data Engineering Solutions Using Microsoft Fabric
Certification	Microsoft Certified: Fabric Data Engineer Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Chinese (Traditional), German, French, Italian, Korean, Spanish, Portuguese (Brazil)
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Annual, free, on Microsoft Learn
Current outline	Skills measured as of October 19, 2026

The three domains

Domain	Weight
Implement and manage an analytics solution	30–35%
Ingest and transform data	30–35%
Monitor and optimize an analytics solution	30–35%

Every domain carries the same range, so each is roughly a third of the exam. That changes how you revise: a candidate who is strong at transformation but has never opened the Monitor hub, configured a deployment pipeline or read an Eventhouse ingestion error will lose about a third of the marks in predictable places.

DP-750 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

DP-750 gives you 120 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Implementing Data Engineering Solutions Using Azure Databricks
Certification	Microsoft Certified: Azure Databricks Data Engineer Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Chinese (Traditional), Korean, German, French, Italian, Portuguese (Brazil), Spanish
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Status	Generally available; the certification page carries no beta label
Current outline	Skills measured as of October 19, 2026

The four domains

Domain	Weight
Set up and configure an Azure Databricks environment	15–20%
Secure and govern Unity Catalog objects	15–20%
Prepare and process data	30–35%
Deploy and maintain data pipelines and workloads	30–35%

Prepare and process data and Deploy and maintain pipelines can together make up 60–70% of the exam. Both are broad. The first covers data modelling choices, six different ingestion routes, transformation, and data quality constraints. The second covers pipeline design, Lakeflow Jobs, Git and testing, bundle deployment, and a long monitoring and troubleshooting section that includes the Spark UI, query profile, OPTIMIZE and VACUUM, and Azure Monitor.

DP-800 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

DP-800 gives you 120 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions or a fixed price. Here is everything else it publishes, and what it means for your preparation.


Exam title	Developing AI-Enabled Database Solutions
Certification	Microsoft Certified: SQL AI Developer Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Chinese (Traditional), Korean, German, French, Italian, Portuguese (Brazil), Spanish
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Annual, free online assessment on Microsoft Learn
Current outline	Skills measured as of October 19, 2026

The three domains

Domain	Weight
Design and develop database solutions	35–40%
Secure, optimize, and deploy database solutions	35–40%
Implement AI capabilities in database solutions	25–30%

Microsoft’s certification page words the third area as “Implement AI capabilities in SQL solutions”; the study guide uses “database solutions”. It is the same area.

Multiple-choice exam tips for IT certifications

Fri, 09 Oct 2026 00:00:00 +0000

To pass a multiple-choice certification exam, read each question for its constraint, remove the options that fail it, and choose the answer the vendor would recommend rather than the one you would improvise at work. Most marks are lost not on missing knowledge but on misreading what was asked. These tips apply to IT and tax exams alike; the vendor-specific rules (timing, review, scoring) you still need to check on the exam’s own page.

PL-300 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

PL-300 gives you 100 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Microsoft Power BI Data Analyst
Certification	Microsoft Certified: Power BI Data Analyst Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Korean, German, French, Spanish, Portuguese (Brazil), Chinese (Traditional), Italian
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Current outline	Skills measured as of April 20, 2026

The four domains

Domain	Weight
Prepare the data	25–30%
Model the data	25–30%
Visualize and analyze the data	25–30%
Manage and secure Power BI	15–20%

The first three domains are equal in weight and together make up the large majority of the exam. Manage and secure Power BI is the smallest, but 15–20% is still a substantial block of questions, and they are some of the most predictable: workspace roles, apps, gateways, scheduled refresh and row-level security.

SC-200 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

SC-200 gives you 100 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Microsoft Security Operations Analyst
Certification	Microsoft Certified: Security Operations Analyst Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Chinese (Traditional), Korean, French, German, Spanish, Portuguese (Brazil), Italian
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Practice assessment	Free, on Microsoft Learn

The three domains

Domain	Weight
Manage a security operations environment	40–45%
Respond to security incidents	35–40%
Perform threat hunting	20–25%

Managing the environment is the largest block and has four groups of objectives: automation for Defender XDR and Sentinel (notifications, Defender for Endpoint settings, attack surface reduction rules, automatic attack disruption, automation rules and playbooks), the Sentinel platform (roles, retention across the analytics, data lake and XDR tiers, workbooks, SOC optimization), data ingestion (connectors, Azure Monitor Agent collection, Syslog and CEF, Windows Event Forwarding, threat indicators, custom tables) and detections (custom detection rules, analytics rules, MITRE ATT&CK coverage, anomalies).

SC-300 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

SC-300 gives you 100 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Microsoft Identity and Access Administrator
Certification	Microsoft Certified: Identity and Access Administrator Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, German, Spanish, French, Italian, Japanese, Korean, Portuguese (Brazil), Chinese (Simplified), Chinese (Traditional)
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Prerequisite certification	None listed
Outline	Skills measured as of October 28, 2026 (English)

The four domains

Domain	Weight
Implement and manage user identities	20–25%
Implement authentication and access management	25–30%
Plan and implement workload identities	20–25%
Plan and automate identity governance	20–25%

Microsoft’s study guide lists the authentication weight as 25–30% in its summary but 20–25% in the section heading, so treat it as the largest area, possibly by only a few points. Authentication and access management can be nearly a third of the exam. It holds the topics that candidates fear most: Conditional Access policy design and troubleshooting, authentication methods and ID Protection. Do not neglect the other three, though. User identities (hybrid sync, external users, administrative units), workload identities (app registrations, enterprise apps, managed identities, Defender for Cloud Apps) and governance (access packages, access reviews, PIM, logs) each carry about a fifth.

SC-401 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

SC-401 gives you 100 minutes, requires a scaled score of 700 out of 1000 to pass, is offered in seven languages and is priced per country at the booking step. Microsoft does not publish the number of questions. This page collects what Microsoft does publish and what it means for how you prepare.


Exam title	Administering Information Security in Microsoft 365
Certification	Microsoft Certified: Information Security Administrator Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Portuguese (Brazil), French, German, Japanese, Chinese (Simplified), Spanish
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Prerequisites	None required; Microsoft expects familiarity with Microsoft 365, PowerShell, Entra, the Defender portal and Defender for Cloud Apps

The three domains

Domain	Weight
Implement information protection	30–35%
Implement data loss prevention and retention	30–35%
Manage risks, alerts, and activities	30–35%

The ranges overlap completely, which means the exam is roughly a third each. In practice the sub-topics are not equal in size. Data loss prevention and retention is two big feature areas sharing one weight, and risks, alerts and activities contains Insider Risk Management, audit, eDiscovery and DSPM for AI. Expect fewer questions per individual feature than the weights suggest, and plan to be competent at all of them rather than expert at one.

SC-900 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 gives you 45 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. This page lists what Microsoft does publish and what it means for your preparation.


Exam title	Microsoft Security, Compliance, and Identity Fundamentals
Certification	Microsoft Certified: Security, Compliance, and Identity Fundamentals
Level	Beginner (fundamentals)
Duration	45 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Korean, French, Spanish, Portuguese (Brazil), Russian, Arabic (Saudi Arabia), Indonesian, German, Chinese (Traditional), Italian
Delivery	Pearson VUE, proctored (Certiport for students and educators)
Price	Set per country or region; shown at booking
Renewal	None; fundamentals certifications do not expire
Current outline	Skills measured as of October 21, 2026

The four skill areas

Skill area	Weight
Describe the concepts of security, compliance, and identity	10–15%
Describe the capabilities of Microsoft Entra	25–30%
Describe the capabilities of Microsoft security solutions	35–40%
Describe the capabilities of Microsoft compliance solutions	20–25%

Security solutions plus Entra make up well over half of the exam. Security solutions is also the area with the most product names to keep apart, so give it the most revision time. The concepts area is the smallest, but its vocabulary (Zero Trust, defense in depth, shared responsibility, authentication versus authorization) is used in the scenarios of the other three, so it pays off beyond its 10–15%.

Subnetting practice trainer

Fri, 09 Oct 2026 00:00:00 +0000

Subnetting is a skill of repetition. This trainer generates new questions without limit, checks your answer at once and shows the working, so you can see which step went wrong. Pick one question type to drill a weak spot or use mixed mode.

How to work a question quickly

Write the prefix as a mask. A /26 leaves 6 host bits, so the last octet changes in steps of 64 (256 − 192).
Find the block: the network address is the largest multiple of the block size that is not above the address.
The broadcast address is the next block start minus 1. The usable range lies between the network and broadcast addresses.
Usable hosts are 2^host bits − 2. A /31 (point-to-point, RFC 3021) has 2 usable addresses and a /32 is a single host.

Worked example. For 192.168.4.77/26 the block size is 64, so the network is 192.168.4.64, the broadcast address is 192.168.4.127, the first host is .65, the last host is .126 and there are 62 usable hosts.

SY0-801 exam format: questions, score and cost

Fri, 09 Oct 2026 00:00:00 +0000

SY0-801 (CompTIA Security+ V8) has a maximum of 90 questions, runs for 90 minutes, and requires 750 on a scale of 100 to 900 to pass. Questions are multiple choice and performance-based. CompTIA’s pages list the launch as expected on or around November 17, 2026, so check the CompTIA site before you book: until then SY0-701 is the exam on offer. Here is what is published, and what is not.

Which IT certification for a career changer

Fri, 09 Oct 2026 00:00:00 +0000

A career changer should start with the certification that matches the destination role, not with the exam that looks easiest. If the destination is undecided, AZ-900 is the safest first step: it has no prerequisites, never expires, and the site’s study plan assumes four weeks at four to six hours a week. If the destination is infrastructure, networking or security, start with Network+ N10-009, and treat Security+ as the second step.

AZ-802 exam format: questions, score and cost

Tue, 29 Sep 2026 00:00:00 +0000

AZ-802 gives you 120 minutes, requires a scaled score of 700 out of 1000 to pass, is offered in English only, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Administering Windows Server
Certification	Microsoft Certified: Windows Server Administrator Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Official practice assessment	Not available yet

The seven domains

Domain	Weight
Deploy and manage AD DS	20–25%
Manage Windows Server instances and workloads in a hybrid environment	10–15%
Manage virtual machines	10–15%
Implement and manage an on-premises and hybrid networking infrastructure	10–15%
Manage storage and file services	15–20%
Secure Windows Server infrastructure	10–15%
Monitor and troubleshoot Windows Server environments	15–20%

Three domains carry most of the weight: AD DS, storage and file services, and monitoring and troubleshooting. Between them they can account for up to 65% of the exam. The other four sit at 10–15% each, which still means several questions apiece. None of them is safe to skip.

AB-900 exam format: questions, score and cost

Mon, 28 Sep 2026 00:00:00 +0000

AB-900 gives you 45 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Microsoft 365 Copilot and Agent Administration Fundamentals
Certification	Microsoft 365 Certified: Copilot and Agent Administration Fundamentals
Level	Fundamentals (beginner)
Duration	45 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian, Spanish
Delivery	Proctored; Pearson VUE, or Certiport for students and educators
Price	Set per country or region; shown at booking
Renewal	None; fundamentals certifications do not expire
Practice assessment	Available free on Microsoft Learn

The three domains

Domain	Weight
Identify the core features and objects of Microsoft 365 services	30–35%
Understand data protection and governance tasks for Microsoft 365 and Copilot	35–40%
Perform basic administrative tasks for Copilot and agents	25–30%

The data protection domain can reach 40% on its own. It is also the domain with the most named tools: Purview Information Protection, DLP, Insider Risk Management, Communication Compliance, DSPM for AI, Data Lifecycle Management, Compliance Manager, Data Explorer, activity explorer, eDiscovery Content search and SharePoint Advanced Management. Expect questions that describe a problem and ask which of those tools finds or fixes it.

AI-300 exam format: questions, score and cost

Mon, 28 Sep 2026 00:00:00 +0000

AI-300 gives you 120 minutes, requires a scaled score of 700 out of 1000 to pass, is offered in English only, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Operationalizing Machine Learning and Generative AI Solutions
Certification	Microsoft Certified: Machine Learning Operations Engineer Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Annually, free, on Microsoft Learn
Replaces	DP-100 (Azure Data Scientist Associate), retired June 1, 2026

The five domains

Domain	Weight
Design and implement an MLOps infrastructure	15–20%
Implement machine learning model lifecycle and operations	25–30%
Design and implement a GenAIOps infrastructure	20–25%
Implement generative AI quality assurance and observability	10–15%
Optimize generative AI systems and model performance	10–15%

Read the table as two halves. The first two domains are about Azure Machine Learning and are worth 40–50% together. The last three are about generative AI in Microsoft Foundry and are worth 40–55%. You cannot pass on one half alone.

AZ-104 exam format: questions, score and cost

Mon, 28 Sep 2026 00:00:00 +0000

AZ-104 gives you 100 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what it means for your preparation.


Exam title	Microsoft Azure Administrator
Certification	Microsoft Certified: Azure Administrator Associate
Level	Intermediate (associate)
Duration	100 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Chinese (Simplified), Chinese (Traditional), French, German, Italian, Japanese, Korean, Portuguese (Brazil), Spanish
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Every 12 months, free, on Microsoft Learn
Current outline	Skills measured as of April 17, 2026

The five domains

Domain	Weight
Manage Azure identities and governance	20–25%
Implement and manage storage	15–20%
Deploy and manage Azure compute resources	20–25%
Implement and manage virtual networking	15–20%
Monitor and maintain Azure resources	10–15%

Identity and compute together can make up half the exam. Compute is also the domain with the most objectives: ARM templates and Bicep, virtual machines, scale sets, containers and App Service all sit inside it. If you only have time to go deep on two areas, pick those two.

AZ-900 exam format: questions, score and cost

Mon, 28 Sep 2026 00:00:00 +0000

AZ-900 gives you 45 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. The certification does not expire. Here is everything else Microsoft publishes, and what it means for your preparation.


Exam title	Microsoft Azure Fundamentals
Certification	Microsoft Certified: Azure Fundamentals
Level	Beginner (fundamentals)
Duration	45 minutes (65 minutes seat time)
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian, Spanish
Delivery	Proctored; booked through Pearson VUE, or through Certiport for students and educators
Price	Set per country or region; shown at booking
Renewal	Not needed; fundamentals certifications do not expire
Outline last updated	July 20, 2026

The three domains

Domain	Weight
Describe cloud concepts	25–30%
Describe Azure architecture and services	35–40%
Describe Azure management and governance	30–35%

The middle domain is the one to plan around. It is the largest, and it holds the most distinct topics: the physical and logical structure of Azure, virtual machines and their alternatives, virtual networking, storage tiers and redundancy, and identity and security. The other two domains are smaller in weight but not small enough to skip; together they are around 60% of the exam.

AI-200 exam format: questions, score and cost

Sun, 27 Sep 2026 00:00:00 +0000

AI-200 gives you 120 minutes, requires a scaled score of 700 out of 1000 to pass, and is priced per country at the booking step. Microsoft does not publish the number of questions. Here is everything else it publishes, and what matters for preparation.


Exam title	Developing AI Cloud Solutions on Azure
Certification	Microsoft Certified: Azure AI Cloud Developer Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian, Spanish
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Annually, free, on Microsoft Learn
Replaces	AZ-204, retired July 31, 2026

The four domains

Domain	Weight
Develop containerized solutions on Azure	20–25%
Develop AI solutions by using Azure data management services	25–30%
Connect to and consume Azure services	20–25%
Secure, monitor, and troubleshoot Azure solutions	20–25%

The spread is fairly flat. The data domain is the only one that can reach 30%, and it is also the one with the most sub-objectives: Cosmos DB for NoSQL, Azure Database for PostgreSQL and Azure Managed Redis each get their own section. If you only have time to go deep on one area, go deep there.

SEE Part 1 format: questions, score and cost

Mon, 21 Sep 2026 00:00:00 +0000

SEE Part 1 (Individuals) has 100 multiple-choice questions, of which 85 are scored, with 3.5 hours of exam time. It is scored on a scale of 200 to 800 and you need 500 to pass. The fee is $317, paid when you schedule with PSI, the IRS’s test vendor. The paper is split into three sections you cannot return to once you have finished them.

That is the short answer. The details below come from the official PSI Candidate Information Bulletin for the July 2026 to February 2027 testing window.

SEE Part 2 format: questions, score and cost

Mon, 21 Sep 2026 00:00:00 +0000

SEE Part 2 is a 100-question multiple-choice exam on business taxation, of which 85 questions are scored and 15 are unscored experimental items. You get 3.5 hours of exam time, the appointment is booked for 4 hours, the fee is $317, and you pass with a scaled score of 500 on a 200–800 scale. Everything below comes from the PSI Candidate Information Bulletin for the July 1, 2026 – February 28, 2027 testing window and the IRS enrolled agent pages.

SEE Part 3 format: questions, score and cost

Mon, 21 Sep 2026 00:00:00 +0000

SEE Part 3 has 100 multiple-choice questions, of which 85 are scored, answered in 3.5 hours. You need a scaled score of 500 on a 200–800 scale, and the fee is $317, paid when you schedule with PSI. The format is identical for all three parts of the Special Enrollment Examination; only the content differs. Part 3 covers representation, practices and procedures — the ethics and process side of being an enrolled agent.

AAIA exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what ISACA publishes about the AAIA exam, and what it leaves out.


Credential	ISACA Advanced in AI Audit (AAIA)
Number of questions	90
Domains	3
Eligibility	CISA or a qualifying audit/accounting designation
Exam fee — ISACA member	US$459
Exam fee — non-member	US$599
Application processing fee	US$50 after passing
Eligibility period after registering	6 months to sit the exam
Window to apply after passing	5 years
Duration	150 minutes (ISACA exam candidate guide)
Passing score	450 on a scaled 200 to 800 range (ISACA exam candidate guide)
Languages	English, Simplified Chinese and Spanish (ISACA exam candidate guide)
Remote proctoring	You can chat with the remote proctor in English only, even if you sit the exam in another language (ISACA exam candidate guide)

Two deadlines, not one

This catches people out. There are two separate clocks:

AAIR exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what ISACA publishes about the AAIR exam.


Credential	ISACA Advanced in AI Risk (AAIR)
Number of questions	90
Domains	3
Eligibility	A recognised designation — see prerequisites
Exam fee — ISACA member	US$459
Exam fee — non-member	US$599
Application processing fee	US$50 after passing
Eligibility period after registering	6 months to sit
Window to apply after passing	5 years
Duration	Not published on ISACA’s AAIR pages
Passing score	Not published on ISACA’s AAIR pages
Languages	English, Simplified Chinese and Spanish (ISACA’s index of exam candidate guides)
Remote proctoring	You can chat with the remote proctor in English only, even if you sit the exam in another language (ISACA exam candidate guide)

Domain weighting

Domain	Weight	Approx. questions
AI Risk Governance and Framework Integration	37%	~33
AI Life Cycle Risk Management	21%	~19
AI Risk Program Management	42%	~38

Counts derive from ISACA’s percentages against the 90-question total.

AAISM exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what ISACA publishes about the AAISM exam, and what it does not.


Credential	ISACA Advanced in AI Security Management (AAISM)
Number of questions	90
Domains	3
Eligibility	Active CISM or CISSP required
Exam fee — ISACA member	US$459
Exam fee — non-member	US$599
Application processing fee	US$50 after passing
Window to apply after passing	5 years
Duration	150 minutes (ISACA exam candidate guide)
Passing score	450 on a scaled 200 to 800 range (ISACA exam candidate guide)
Languages	English, Spanish and Japanese (ISACA exam candidate guide)
Remote proctoring	You can chat with the remote proctor in English only, even if you sit the exam in another language (ISACA exam candidate guide)

The prerequisite is a gate, not advice

Most certifications “recommend” experience. ISACA states plainly that AAISM candidates must hold an active CISM or CISSP certification. If you hold neither, AAISM is not available to you yet, and no amount of AI knowledge substitutes.

AB-620 exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Microsoft publishes about the AB-620 format, and what it leaves unstated.


Exam title	Designing and Building Integrated AI Solutions in Copilot Studio
Certification	Microsoft Certified: AI Agent Builder Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Annually, free, on Microsoft Learn

Languages: English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian and Spanish. If your language is not offered, you can request an extra 30 minutes.

Agentforce Specialist exam format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Salesforce publishes about the Agentforce Specialist exam.


Certification	Salesforce Certified Agentforce Specialist
Scored questions	60 multiple choice
Unscored questions	Up to 5, randomly integrated
Duration	105 minutes
Passing score	72%
Registration fee	USD 200 (JPY 30,000), plus local taxes
Retake fee	USD 100 (JPY 15,000), plus local taxes
Prerequisite	None
Languages	English, French, Japanese in the official exam guide. A Salesforce announcement on Trailhead lists this exam among 13 now available in Latin American Spanish, but the exam guide does not confirm it, so check Salesforce’s language options page when you book
Release alignment	Spring ‘26
Delivery	Test centre or online proctored
References	No hard-copy or online materials permitted
Maintenance	Annual Trailhead module required

A real percentage, not a scaled score

Unusually among the certifications covered here, Salesforce publishes a plain 72% rather than a scaled score on some arbitrary range. With 60 scored questions, that means you need 44 correct.

AI-103 exam format: questions, passing score, time and cost

Sun, 20 Sep 2026 00:00:00 +0000

Knowing the format removes a lot of exam-day stress. Here are the facts for AI-103.


Duration	120 minutes
Number of questions	Not published by Microsoft
Passing score	700 on a scale of 1–1000
Price	$165 in the US (varies by country)
Languages	English, Chinese (Simplified), Chinese (Traditional), French, German, Italian, Japanese, Korean, Portuguese (Brazil), Spanish
Delivery	Pearson VUE test centre or online proctored
Validity	One year, renewable for free on Microsoft Learn

Question types

Microsoft exams mix several formats: single and multiple choice, drag-and-drop ordering, case studies with several questions about one scenario, and yes/no statement series. For the yes/no series you usually cannot go back once you have answered, so read them carefully.

AI-901 exam format: questions, passing score, time and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Microsoft actually publishes about the AI-901 format, and what it does not. Knowing the difference matters, because a lot of what circulates online about fundamentals exams is guesswork.


Certification earned	Microsoft Certified: Azure AI Fundamentals
Level	Fundamentals
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Duration	Not published on the exam page
Price	Set per country or region; shown when you book
Delivery	Pearson VUE, test centre or online proctored
Retirement date	None announced

Languages: English, Arabic (Saudi Arabia), Chinese (Simplified), Chinese (Traditional), French, German, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), Russian and Spanish. If the exam is not offered in your preferred language, you can request an extra 30 minutes.

AIF-C01 exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

AWS publishes more detail about this exam than most vendors do. Here are the facts.


Exam code	AIF-C01
Certification	AWS Certified AI Practitioner
Level	Foundational
Languages	English, Arabic, French, Japanese, Korean, Portuguese (Brazil), Spanish (Latin America and Spain), Chinese (Simplified and Traditional); German and Italian only until October 15, 2026 (AWS exam page)
Duration	90 minutes
Questions	65 total — 50 scored, 15 unscored
Passing score	700 on a scaled range of 100–1,000
Price	100 USD
Scoring	Compensatory — you pass overall, not per domain
Penalty for guessing	None

Fifteen questions do not count

AWS includes 15 unscored questions that do not affect your result. They are collecting data on candidate questions for future use, and they are not identified on the exam.

AIF-C01 vs AI-901

Sun, 20 Sep 2026 00:00:00 +0000

Take AIF-C01 if your organisation runs on AWS, AI-901 if it runs on Azure. Both are foundational, both are conceptual rather than hands-on, and neither will impress an engineer. The deciding factor is which cloud you actually work near — not which exam is better, because they are close to equivalent in difficulty and value.

Side by side

	AIF-C01	AI-901
Vendor	AWS	Microsoft
Level	Foundational	Fundamentals
Domains	5	2
Heaviest domain	Applications of Foundation Models (28%)	Implement AI solutions with Microsoft Foundry (55–60%)
Coding required	No	No
Prerequisite	None	None

Where they differ in substance

AIF-C01 spreads wider. Its five domains cover AI and ML fundamentals, generative AI fundamentals, applications of foundation models, responsible AI, and security, compliance and governance. Roughly a quarter of the exam is governance and responsible AI, which is more than most people expect from a foundational certificate.

Azure AI certification path

Sun, 20 Sep 2026 00:00:00 +0000

Start with AI-901 if you are new to AI on Azure, then move to AI-103 if you build AI apps and agents in code, AI-200 if you build the back-end services they run on, AI-300 if you run models and generative AI in production, or AB-620 if you build agents without writing code. Microsoft 365 administrators who govern Copilot start with AB-900 instead. SC-500 is the security specialisation and assumes you already work in cloud security. There is no need to collect them all.

CY0-001 exam format: questions, passing score, time and cost

Sun, 20 Sep 2026 00:00:00 +0000

CY0-001 is a short exam with an unusual amount packed into it. Sixty minutes, up to sixty questions, and a mix of multiple choice and performance-based items. Here are the published facts.


Exam code	CY0-001
Certification	CompTIA SecAI+
Version	V1
Duration	60 minutes
Number of questions	Maximum of 60
Question types	Multiple choice and performance-based
Passing score	600 on a scale of 100–900
Languages	English and Japanese
Launched	February 17, 2026
Estimated retirement	Around three years after launch
Price	Not published on the certification page; check at booking

One minute per question

That ratio is the defining feature of this exam. Sixty questions in sixty minutes leaves no room for extended deliberation, and performance-based items eat more than their share of the clock.

Databricks Data Analyst Associate format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Data Analyst Associate exam.


Certification	Databricks Certified Data Analyst Associate
Scored questions	45
Duration	90 minutes
Question format	Multiple choice
Registration fee	$200
Validity	2 years, then recertification on the current version
Prerequisites	None (training strongly recommended)
Recommended experience	6+ months of hands-on data analysis tasks
Delivery	Online proctored or test centre
Languages	English only
Test aids	None permitted
Passing score	Not published by Databricks

English only

Worth noting because it differs from Databricks’ other certifications, which are offered in English, Japanese, Portuguese (Brazil) and Korean. This one is English only, so if that is not your first language, factor that into your preparation.

Databricks Data Engineer Associate format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Data Engineer Associate exam.


Certification	Databricks Certified Data Engineer Associate
Scored questions	45
Duration	90 minutes
Question format	Multiple choice
Registration fee	$200
Validity	2 years, then recertification
Prerequisites	None
Recommended experience	Hands-on with the exam guide’s data engineering tasks
Delivery	Online proctored or test centre
Languages	English, Japanese, Portuguese (BR), Korean
Test aids	None permitted
Passing score	Not published by Databricks

Databricks notes that unscored items are included for statistical purposes and do not affect your score, so the paper may run longer than 45 questions.

Databricks Data Engineer Professional format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Data Engineer Professional exam. The figures are for the version that went live on 9 October 2026.


Certification	Databricks Certified Data Engineer Professional
Scored questions	60
Unscored questions	Up to 10 may be included, unmarked; extra time is already in the limit
Duration	120 minutes
Question format	Multiple choice
Registration fee	$200, plus local taxes
Validity	2 years, then recertification
Prerequisites	None (related training highly recommended)
Recommended experience	One year hands-on with the exam guide’s data engineering tasks
Delivery	Online proctored or test centre
Languages	English, Japanese, Portuguese (Brazil), Korean (Databricks’ exam page; it does not say whether the October 9 outline is localized yet)
Test aids	None permitted
Passing score	Not published by Databricks

The outline changed on 9 October 2026

Databricks replaced the outline on 9 October 2026. Anyone who sat the exam up to 8 October took the old version: 59 scored questions across ten sections, offered in English, Japanese, Portuguese (Brazil) and Korean. From 9 October the exam has 60 scored questions across nine sections. Databricks’ exam page still lists the same four languages and does not say whether the new outline is localized yet. Time, fee and validity did not change.

Databricks GenAI Engineer exam format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Generative AI Engineer Associate exam.


Certification	Databricks Certified Generative AI Engineer Associate
Scored questions	45
Duration	90 minutes
Question format	Multiple choice only
Registration fee	$200
Validity	2 years, then recertification
Prerequisites	None, though training is strongly encouraged
Recommended experience	6+ months hands-on with generative AI tasks
Delivery	Online proctored or test centre
Languages	English, Japanese, Portuguese (BR), Korean
Test aids	None permitted
Passing score	Not published by Databricks

Two minutes per question

45 scored questions in 90 minutes gives exactly two minutes each — comfortable by certification standards. Databricks notes that unscored items may be included, so the paper you sit may run longer than 45 questions. Pace yourself against the clock rather than the question count.

Databricks ML Associate exam format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Machine Learning Associate exam.


Certification	Databricks Certified Machine Learning Associate
Scored questions	48
Duration	90 minutes
Question format	Multiple choice
Registration fee	$200
Validity	2 years, then recertification
Prerequisites	None, though related training is recommended
Recommended experience	6+ months hands-on with the exam’s ML tasks
Delivery	Online proctored or test centre
Languages	English, Japanese, Portuguese (BR), Korean
Test aids	None permitted
Passing score	Not published by Databricks

Under two minutes per question

48 scored questions in 90 minutes is about 112 seconds each. Slightly tighter than the Generative AI Engineer exam, which allows two minutes. Not pressured, but not leisurely either.

Databricks ML Professional exam format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Machine Learning Professional exam.


Certification	Databricks Certified Machine Learning Professional
Scored questions	59
Duration	120 minutes
Question format	Multiple choice
Registration fee	$200
Validity	2 years, then recertification on the current version
Prerequisites	None (training highly recommended)
Recommended experience	1+ years hands-on with the exam guide’s ML tasks
Delivery	Online proctored or test centre
Languages	English only
Test aids	None permitted
Passing score	Not published by Databricks

Databricks notes that unscored items may be included for statistical purposes and are not identified to candidates.

MLA-C02 exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

MLA-C02 is currently offered under beta conditions, which changes several published figures. Here is what AWS states, and where the sources differ.


Exam code	MLA-C02
Certification	AWS Certified Machine Learning Engineer – Associate
Level	Associate
Languages	Beta: English only. AWS plans Japanese, Korean and Simplified Chinese for the standard exam and does not list Spanish; MLA-C01 offered English, Japanese, Korean and Simplified Chinese (AWS exam page)
Duration	170 minutes (beta)
Questions	85 (beta)
Price	75 USD (beta pricing)
Passing score	720 on a scaled range of 100–1,000
Scoring	Compensatory — you pass overall, not per domain
Validity	3 years
Penalty for guessing	None

The beta discrepancy, explained

AWS’s certification page lists 85 questions in 170 minutes at 75 USD. The official exam guide describes 50 scored questions plus 15 unscored, which totals 65.

N10-009 exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what CompTIA publishes about Network+ N10-009.


Exam code	N10-009
Certification	CompTIA Network+ (V9)
Number of questions	Maximum 90
Question types	Multiple choice and performance-based
Duration	90 minutes
Passing score	720 on a scale of 100–900
Launched	June 20, 2024
Estimated retirement	Around 2027
Languages	English, German, Japanese, Portuguese, Spanish
Recommended	A+ plus 9–12 months hands-on experience
Price	Not listed on CompTIA’s certification page; check at booking

One minute per question

Ninety questions in ninety minutes, and performance-based items take far longer than multiple choice. That is the defining constraint of this exam.

SC-500 exam format: questions, score and cost

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Microsoft publishes about the SC-500 format, and the two details that change how you should prepare.


Exam title	Implementing End-to-End Security Controls for Cloud and AI Workloads
Certification	Microsoft Certified: Cloud and AI Security Engineer Associate
Level	Intermediate (associate)
Duration	120 minutes
Passing score	700 on a scale of 1–1000
Number of questions	Not published by Microsoft
Languages	English, Japanese, Chinese (Simplified), Chinese (Traditional), Korean, German, French, Italian, Portuguese (Brazil), Spanish
Delivery	Pearson VUE, proctored
Price	Set per country or region; shown at booking
Renewal	Annually, free, on Microsoft Learn

Languages

Microsoft lists ten languages for SC-500, the same number as the retired AZ-500 had. The English version is always updated first and localised versions follow later, so shortly after an outline change you may get the previous outline in another language. If the exam is not offered in your language you can request an additional 30 minutes; ask before booking, not on the day.

Spark Developer Associate exam format

Sun, 20 Sep 2026 00:00:00 +0000

Here is what Databricks publishes about the Associate Developer for Apache Spark exam.


Certification	Databricks Certified Associate Developer for Apache Spark
Scored questions	45
Duration	90 minutes
Question format	Multiple choice
Registration fee	$200
Programming language	Python
Validity	2 years, then recertification on the current version
Prerequisites	None (training recommended)
Recommended experience	6+ months hands-on practice
Delivery	Online proctored or test centre
Languages	English only
Test aids	None permitted
Passing score	Not published by Databricks

Unscored items may be included for statistical purposes without affecting your score.

200-901 vs CCNA 200-301: which exam first?

Fri, 09 Oct 2026 00:00:00 +0000

Take the CCNA 200-301 first if you work on or want to work on networks, and take 200-901 first if you are a developer or scripter who needs Cisco platform and API skills. They test different halves of the same job. 200-301 asks whether you can build, verify and troubleshoot a network from the command line. 200-901 asks whether you can automate and integrate with it using code, REST APIs and Cisco’s platforms. Neither is a prerequisite for the other.

AB-100 vs AB-620: architect or builder?

Fri, 09 Oct 2026 00:00:00 +0000

Take AB-620 if you build agents in Copilot Studio, and AB-100 if you decide how agents are used across a business. AB-620 is an associate-level builder exam; AB-100 is an expert-level architect exam. They are also connected: AB-620 is one of the associate certifications Microsoft accepts as the required prerequisite for the AB-100 credential. For many people they are two steps of one route, not alternatives.

Side by side

	AB-100	AB-620
Certification	Agentic AI Business Solutions Architect Expert	AI Agent Builder Associate
Level	Expert	Associate
Main verb	Plan, design, govern	Configure, integrate, test
Domains	Plan 25–30%, Design 25–30%, Deploy 40–45%	Plan and configure 30–35%, Integrate and extend 40–45%, Test and manage 20–25%
Platforms	Copilot Studio, Microsoft 365 Copilot, Microsoft Foundry, Dynamics 365, Power Platform	Copilot Studio with Power Platform and Foundry
Typical question	Which approach fits this business constraint?	How do you configure this agent to meet this requirement?
Needs another certification	Yes, one qualifying associate certification	No

How the content differs

AB-620 goes deep on one tool. It covers topics, knowledge sources, connectors, MCP tools, computer use and Power Platform ALM inside Copilot Studio. You are expected to know how it is done.

AB-730 vs AB-731: which should you take?

Fri, 09 Oct 2026 00:00:00 +0000

Take AB-730 if you use AI tools yourself and want to prove you can do it well; take AB-731 if you decide where and how your organisation adopts AI. AB-730, AI Business Professional, is about the individual user: prompts, chats, documents, meetings and agents inside Microsoft 365 Copilot. AB-731, AI Transformation Leader, is about the decision-maker: business value, choosing Microsoft AI services, responsible AI governance and adoption planning. Neither requires you to write code, and both are 45-minute exams.

AB-731 vs AB-730: leader or practitioner?

Fri, 09 Oct 2026 00:00:00 +0000

Take AB-731 if your job is deciding where AI goes, who owns it and how it is governed; take AB-730 if your job is using Copilot well every day. AB-731 (AI Transformation Leader) is a strategy exam about choosing and rolling out Microsoft AI. AB-730 (AI Business Professional) is a practitioner exam about prompting, drafting and analysing with generative AI. They share vocabulary, they are both 45-minute business exams that need no coding, and you can sit them in either order.

AZ-305 vs AZ-400: which expert exam first?

Fri, 09 Oct 2026 00:00:00 +0000

Take AZ-305 first if you design or advise on Azure infrastructure, and AZ-400 first if you spend your days on pipelines, source control and releases. Both are expert-level Microsoft exams, both build on the Azure Administrator Associate certification, and neither requires the other. AZ-305 asks which Azure services meet a set of business requirements; AZ-400 asks how to build and automate the delivery of software to them.

The two exams side by side

	AZ-305	AZ-400
Exam title	Designing Microsoft Azure Infrastructure Solutions	Designing and Implementing Microsoft DevOps Solutions
Certification	Azure Solutions Architect Expert	DevOps Engineer Expert
Level	Expert	Expert
Prerequisite certification	Azure Administrator Associate	Azure Administrator Associate (the Azure Developer Associate option is retired)
Passing score	700	700
Languages offered	10	10
Renewal	Free, annually	Free, annually
Main verb in the objectives	Recommend, design, evaluate	Design, implement, configure

The prerequisite difference matters. If you hold AZ-104, both certifications are open to you. Microsoft’s certification page still names the Azure Developer Associate (exam AZ-204) as an alternative for AZ-400, but marks that certification and its renewal assessment as retired, so new candidates should plan on AZ-104. The architect credential has always required the administrator certification.

AZ-400 vs AZ-104: which exam first?

Fri, 09 Oct 2026 00:00:00 +0000

Take AZ-104 first unless you already hold the Azure Administrator Associate certification. AZ-104 tests whether you can configure and operate Azure resources; AZ-400 tests whether you can design and run the delivery process around them with GitHub, Azure DevOps and pipelines. Microsoft’s DevOps Engineer Expert page requires AZ-400 plus an associate certification, and for a new candidate the administrator certification, earned with AZ-104, is the practical one.

The two exams side by side

	AZ-104	AZ-400
Exam title	Microsoft Azure Administrator	Designing and Implementing Microsoft DevOps Solutions
Certification	Azure Administrator Associate	DevOps Engineer Expert
Level	Associate (intermediate)	Expert
Prerequisite certification	None	Azure Administrator Associate or Azure Developer Associate (the latter is retired)
Passing score	700	700
Languages offered	10	10
Renewal	Free, every 12 months	Free, annually
Main verb in the objectives	Configure, create, manage	Design, implement, configure

Domains compared

AZ-104 domain	Weight	AZ-400 domain	Weight
Manage Azure identities and governance	20-25%	Develop a security and compliance plan	10-15%
Deploy and manage Azure compute resources	20-25%	Design and implement build and release pipelines	50-55%
Implement and manage storage	15-20%	Design and implement a source control strategy	10-15%
Implement and manage virtual networking	15-20%	Design and implement processes and communications	10-15%
Monitor and maintain Azure resources	10-15%	Implement an instrumentation strategy	5-10%

The rows are not equivalents; they only line up by weight order. AZ-104 spreads its weight across five Azure areas. AZ-400 puts more than half on pipelines.

CCNA 200-301 vs Network+ N10-009

Fri, 09 Oct 2026 00:00:00 +0000

Take Network+ first if you are new to networking and want a vendor-neutral foundation; take the CCNA first if you already know the basics and want a credential that proves you can configure and troubleshoot switches and routers. Network+ asks you to understand and diagnose networks across vendors. The CCNA asks you to build and verify them on Cisco platforms, which is a harder, more command-driven exam.

The two exams side by side

	CCNA 200-301	Network+ N10-009
Issuer	Cisco	CompTIA
Level	Associate	Entry to intermediate
Duration	120 minutes	90 minutes
Questions	Not published	Maximum 90
Passing score	Not published	720 on a scale of 100-900
Vendor focus	Cisco IOS XE style configuration	Vendor-neutral
Prerequisites	None required	None required (A+ and 9-12 months recommended)
Validity	3 years	See the Network+ guide

Both are open to anyone. The practical difference is the expected skill: a Network+ candidate should explain what a VLAN, OSPF or NAT is and spot the cause of a fault; a CCNA candidate should also write the configuration and interpret the show output on a Cisco device.

Certification exam question types explained

Fri, 09 Oct 2026 00:00:00 +0000

IT certification exams use several question formats: single-answer multiple choice, multiple response, drag-and-drop or ordering, dropdown and hot-area items, long scenario or case-study sets, and hands-on or performance-based tasks. Which ones you meet depends on the vendor and the exam, so read the exam page before you book. This guide explains each format and how to approach it.

Single answer

One correct option out of three to five. These make up the bulk of most exams. The method is in the multiple-choice tips: find the constraint, remove the options that fail it, choose what the vendor recommends.

DP-600 vs DP-700: analytics or data engineer?

Fri, 09 Oct 2026 00:00:00 +0000

Take DP-600 if your work is modelling, securing and serving data for analysis, with SQL, DAX and semantic models. Take DP-700 if your work is ingesting, transforming and orchestrating data with pipelines, notebooks and PySpark. Both are associate-level Microsoft Fabric exams of 100 minutes with no prerequisite between them, so you choose by role, not by order.

The two certifications side by side

	DP-600	DP-700
Certification	Fabric Analytics Engineer Associate	Fabric Data Engineer Associate
Duration	100 minutes	100 minutes
Level	Intermediate	Intermediate
Focus	Analytical assets: semantic models, warehouses, lakehouses	Data loading patterns, architectures, orchestration
Main languages	SQL, KQL, DAX	SQL, PySpark, KQL
Role (per Microsoft)	Data Engineer, Data Analyst	Data Engineer

Both certifications share the 12-month renewal on Microsoft Learn. The shared word is KQL: both exams expect you to query real-time data. The difference is the second and third language. DP-600 adds DAX, the language of semantic models. DP-700 adds PySpark, the language of notebook-based transformation.

DP-700 vs Databricks Data Engineer Associate

Fri, 09 Oct 2026 00:00:00 +0000

Take DP-700 if your data work runs on Microsoft Fabric, and the Databricks Data Engineer Associate if it runs on Databricks. The two exams certify the same job on different platforms, and neither is a substitute for the other. DP-700 is broader across a whole analytics platform (workspaces, security, orchestration, streaming and monitoring); the Databricks exam is narrower and goes deeper into Spark, Delta Lake and Lakeflow Jobs.

The two exams side by side

	DP-700	Databricks Data Engineer Associate
Vendor credential	Microsoft Certified: Fabric Data Engineer Associate	Databricks Certified Data Engineer Associate
Duration	100 minutes	90 minutes
Scored questions	Not published	45
Passing score	700 on a 1–1000 scale	Not published
Registration fee	Set per country or region	$200
Validity	12 months, free renewal assessment	2 years, then recertification
Languages	10	4
Prerequisites	None required	None

The Databricks figures come from that exam’s own format page. Only DP-700 publishes a passing score, which gives you a concrete target.

DP-750 vs DP-700: Databricks or Fabric?

Fri, 09 Oct 2026 00:00:00 +0000

Take DP-750 if your pipelines run on Azure Databricks, and DP-700 if they run on Microsoft Fabric. Both are Microsoft associate-level data engineering exams with a passing score of 700, so neither is an easier shortcut to the other. They test different platforms, and the platform you work on should decide the choice.

The two exams side by side

	DP-750	DP-700
Exam title	Implementing Data Engineering Solutions Using Azure Databricks	Implementing Data Engineering Solutions Using Microsoft Fabric
Certification	Azure Databricks Data Engineer Associate	Fabric Data Engineer Associate
Platform	Azure Databricks: Spark clusters and serverless compute, Unity Catalog, Lakeflow	Microsoft Fabric: lakehouses, warehouses, pipelines, eventhouses
Languages named in the role	SQL and Python	SQL, PySpark and KQL
Passing score	700 / 1000	700 / 1000
Duration	120 minutes	See the DP-700 section
Skill areas	Four	Three

What the skill areas look like

DP-750 splits into setting up the environment, securing and governing Unity Catalog, preparing and processing data, and deploying and maintaining pipelines. Compute selection, Unity Catalog privileges, Spark Structured Streaming, Auto Loader, Lakeflow Jobs and bundle deployment all belong to Databricks.

DP-800 vs DP-700: which exam fits you?

Fri, 09 Oct 2026 00:00:00 +0000

Take DP-800 if you write T-SQL and build applications on top of a SQL database; take DP-700 if you build data pipelines and lakehouses in Microsoft Fabric. DP-800 is a developer exam about the database itself: schema, programmability, security, deployment and AI search inside SQL. DP-700 is a data engineering exam about moving and transforming data at scale with pipelines, notebooks and KQL. Neither is a prerequisite for the other.

PL-300 vs DP-600: which exam first?

Fri, 09 Oct 2026 00:00:00 +0000

Take PL-300 if your job is building reports and semantic models for business users in Power BI; take DP-600 if you design and manage the analytics platform underneath them in Microsoft Fabric. PL-300 is the analyst exam: Power Query, DAX, report design and the Power BI service. DP-600 is the Fabric analytics engineer exam: lakehouses, warehouses, SQL, KQL, DAX and enterprise-scale semantic models. Neither certification page lists the other as a prerequisite, so you can start with either, but most people who need both take PL-300 first.

Risk calculation drill: SLE, ALE, ARO

Fri, 09 Oct 2026 00:00:00 +0000

Quantitative risk questions are pure arithmetic once you know which formula applies. This drill generates fresh numbers each time and shows the calculation after every answer. The calculator below lets you check your own scenario.

The formulas

SLE (single loss expectancy) = asset value × exposure factor (EF).
ALE (annualized loss expectancy) = SLE × ARO (annual rate of occurrence).
Control benefit = ALE before − ALE after − annual cost of the control. A positive result means the control pays for itself.

Worked example. A server worth $100,000 loses 25% of its value in one incident, so the SLE is $25,000. If the incident is expected twice a year, the ALE is $50,000. A control that cuts the rate to 0.5 a year leaves an ALE of $12,500, saving $37,500 a year before its own cost.

SC-200 vs SC-500: analyst or engineer?

Fri, 09 Oct 2026 00:00:00 +0000

Choose SC-200 if your job is watching, investigating and responding: triaging incidents in Defender XDR, writing detections in Sentinel and hunting with KQL. Choose SC-500 if your job is building the protections in the first place: securing identities, storage, networks and compute in Azure and hybrid environments, including AI workloads. Neither is a prerequisite for the other, and many security careers use both.

The two exams side by side

	SC-200	SC-500
Exam title	Microsoft Security Operations Analyst	Implementing End-to-End Security Controls for Cloud and AI Workloads
Certification	Security Operations Analyst Associate	Cloud and AI Security Engineer Associate
Level	Associate (intermediate)	Associate (intermediate)
Duration	100 minutes	120 minutes
Passing score	700	700
Role	Security operations analyst	Security engineer
Core tools	Defender XDR, Sentinel, KQL	Entra, Defender for Cloud, Azure networking, storage and compute security
Renewal	Free, every 12 months	Free, annually

Domains compared

SC-200	Weight	SC-500	Weight
Manage a security operations environment	40–45%	Manage identity, access, and governance	20–25%
Respond to security incidents	35–40%	Secure storage, databases, and networking	25–30%
Perform threat hunting	20–25%	Secure compute	20–25%
		Manage and monitor security posture	20–25%

SC-200 is lopsided towards two areas; SC-500 is almost flat. That changes how you revise: on SC-200 you can concentrate on the big two plus KQL, while SC-500 punishes any blind spot.

SC-300 vs SC-401: which exam fits you?

Fri, 09 Oct 2026 00:00:00 +0000

Take SC-300 if your job is who can sign in and what they can reach; take SC-401 if your job is protecting the data once they are in. SC-300 leads to the Identity and Access Administrator Associate credential and centres on Microsoft Entra. SC-401 leads to the Information Security Administrator Associate credential and centres on Microsoft Purview. Neither is a prerequisite for the other, and Microsoft lists no prerequisite certification for either.

SC-401 vs SC-200: which exam to take?

Fri, 09 Oct 2026 00:00:00 +0000

Take SC-401 if your job is protecting data: labels, DLP, retention, insider risk and audit in Microsoft Purview. Take SC-200 if your job is defending against attacks: triaging incidents, hunting threats and writing detections in Microsoft Defender XDR and Microsoft Sentinel. They are both Microsoft security associate exams of 100 minutes, but they sit in different parts of the Microsoft security stack and share very little content.

The two exams side by side

	SC-401	SC-200
Certification	Information Security Administrator Associate	Security Operations Analyst Associate
Focus	Protect and govern sensitive data	Detect, investigate and respond to threats
Main tools	Microsoft Purview, Defender for Cloud Apps	Microsoft Defender XDR, Microsoft Sentinel, Defender for Cloud
Domains	Information protection; DLP and retention; risks, alerts and activities	Manage a security operations environment; respond to security incidents; perform threat hunting
Duration	100 minutes	100 minutes
Passing score	700	700
Languages	7	10
Renewal	Free, every 12 months	Free, every 12 months

Who does what

The information security administrator decides which documents are confidential, applies encryption through sensitivity labels, writes policies that block a payroll file from being emailed out, sets retention for records and investigates risky behaviour by employees. The work is policy design and tuning.

SC-900 vs AZ-900: which exam first?

Fri, 09 Oct 2026 00:00:00 +0000

Take AZ-900 first if you want a general cloud foundation or plan to work with Azure infrastructure; take SC-900 first if your job or next step is security, identity or compliance. Neither is a prerequisite for the other, and Microsoft requires no fundamentals exam before any role-based exam. The choice is about what you want to learn first and where you are heading.

The two exams side by side

	SC-900	AZ-900
Subject	Security, compliance and identity concepts and Microsoft products	Cloud concepts, Azure services, governance and cost
Duration	45 minutes	45 minutes
Passing score	700 / 1000	700 / 1000
Level	Fundamentals	Fundamentals
Biggest area	Security solutions, 35–40%	Architecture and services, 35–40%
Renewal	Does not expire	Does not expire

Both are conceptual, both last 45 minutes, and neither needs a subscription or any code. See the AZ-900 hub for the other side.

Study for IT certifications with a full-time job

Fri, 09 Oct 2026 00:00:00 +0000

Studying for an IT certification while working full time is a scheduling problem more than a motivation problem: take the total hours the exam’s study plan implies, divide by the hours you can protect each week, and book a date that arithmetic supports. For most fundamentals exams on this site that is 16 to 28 hours in total, so four to six weeks at a few hours a week; for associate-level exams it is roughly 32 to 40 hours for someone who already works with the technology.

SY0-801 vs SY0-701: which Security+ to take

Fri, 09 Oct 2026 00:00:00 +0000

SY0-801 is the next version of SY0-701, not a different certification. Both are CompTIA Security+, with the same format (up to 90 questions, 90 minutes, 750 on a 100-900 scale) and five domains. The differences are in the objectives: V8 adds a dedicated AI objective, shifts the domain weights, and updates the topic lists. The rule of thumb is simple. If you are ready before SY0-801 is available, or well into SY0-701 material, sit SY0-701. If you are starting from zero now, study for SY0-801.

AZ-802 vs AZ-800 and AZ-801

Tue, 29 Sep 2026 00:00:00 +0000

AZ-802 replaces both AZ-800 and AZ-801. From 30 September 2026, when Microsoft retires the two older exams, AZ-802 is the only exam that earns the Windows Server administrator certification. It is not a merger of everything the pair covered: Microsoft kept the day-to-day administration and security topics, and dropped containers, migration, clustering as a separate domain, backup and disaster recovery, and hybrid identity synchronisation.

The exams side by side

	AZ-800	AZ-801	AZ-802
Exam title	Administering Windows Server Hybrid Core Infrastructure	Configuring Windows Server Hybrid Advanced Services	Administering Windows Server
Status	Retires 30 September 2026	Retires 30 September 2026	Current
Earns the certification	Only together with AZ-801	Only together with AZ-800	On its own
Domains	5	5	7
Passing score	700	700	700
Languages	7	10	English only

The certification is being renamed at the same time: Windows Server Hybrid Administrator Associate becomes Windows Server Administrator Associate.

AB-900 vs MS-900: what changed?

Mon, 28 Sep 2026 00:00:00 +0000

MS-900, Microsoft 365 Fundamentals, retired on March 31, 2026. AB-900 is the fundamentals exam that people coming from MS-900 now look at, but it is not the same exam with a new number. MS-900 asked you to describe Microsoft 365: cloud concepts, apps, security and licensing. AB-900 asks you to identify, monitor and perform basic administrative work in a tenant where Copilot and agents are running. It is narrower in scope and noticeably more hands-on.

AI-300 vs DP-100: what changed?

Mon, 28 Sep 2026 00:00:00 +0000

AI-300 replaced DP-100 when the Azure Data Scientist Associate certification retired on June 1, 2026. Much of the Azure Machine Learning content carries over — workspaces, compute, MLflow, pipelines, endpoints — but the exam now tests operating models rather than building them, and nearly half of it is about generative AI in Microsoft Foundry. If you studied for DP-100, you have a solid base for two domains and little for the other three.

AZ-104 vs AZ-305: which exam first?

Mon, 28 Sep 2026 00:00:00 +0000

Take AZ-104 first. AZ-104 tests whether you can configure and operate Azure resources; AZ-305 tests whether you can design a solution from business requirements. Microsoft makes the order official: you must hold the Azure Administrator Associate certification, earned by passing AZ-104, before passing AZ-305 earns you the Azure Solutions Architect Expert credential.

The two exams side by side

	AZ-104	AZ-305
Exam title	Microsoft Azure Administrator	Designing Microsoft Azure Infrastructure Solutions
Certification	Azure Administrator Associate	Azure Solutions Architect Expert
Level	Associate (intermediate)	Expert
Prerequisite certification	None	Azure Administrator Associate
Passing score	700	700
Languages offered	10	10
Verb in the objectives	Configure, create, manage	Recommend, design, evaluate
Renewal	Free, every 12 months	Free, annually

You can sit AZ-305 before AZ-104; nothing stops you from booking it. But the expert certification is only awarded once both are in place, and AZ-305 assumes the hands-on knowledge that AZ-104 tests.

AZ-900 vs AI-901: which to take first?

Mon, 28 Sep 2026 00:00:00 +0000

AZ-900 and AI-901 are both Microsoft fundamentals exams, but they test different things. AZ-900 covers the cloud itself: what Azure is, how it is structured, what the core services do and how it is secured, governed and billed. AI-901 covers building small AI solutions on Azure with Microsoft Foundry. For most people new to Azure, AZ-900 comes first, because AI-901 assumes the vocabulary AZ-900 teaches.

The two exams side by side

	AZ-900	AI-901
Certification	Microsoft Certified: Azure Fundamentals	Microsoft Certified: Azure AI Fundamentals
Level	Fundamentals	Fundamentals
Passing score	700 / 1000	700 / 1000
Skill areas	Three	Two
Style	Conceptual: describe, define, compare	Mixed: concepts plus hands-on implementation
Coding	None	Python syntax and SDK familiarity expected
Expires	No	No
Languages	13	13

Both are fundamentals certifications, so neither has to be renewed.

AI-200 vs AZ-204: what changed?

Sun, 27 Sep 2026 00:00:00 +0000

AI-200 replaced AZ-204 when the Azure Developer Associate certification retired on July 31, 2026. About half of what AZ-204 tested carries over — Functions, Service Bus, Event Grid, Key Vault, App Configuration, containers — but the storage domain has been rebuilt around vector databases, and Python is now the named language. If you studied for AZ-204, you are not starting from zero, but you are not close to done either.

SEE Part 1 vs Part 2: which to take first

Mon, 21 Sep 2026 00:00:00 +0000

For most candidates, take Part 1 first. Part 1 covers individual income tax, and Part 2 builds on it: a partnership’s income ends up on a partner’s individual return, basis rules are the same idea at both levels, and self-employment tax appears in both. The IRS lets you take the three parts in any order, so this is a strategy choice, not a rule. The exception is someone who prepares business returns every day and rarely sees a Form 1040.

SEE Part 2: Enrolled Agent vs CPA

Mon, 21 Sep 2026 00:00:00 +0000

For representing clients before the IRS, an Enrolled Agent and a CPA have the same rights: the IRS states that enrolled agents, CPAs and attorneys all have unlimited representation rights, covering any taxpayer, any tax matter and any IRS office. The difference is everything else. The EA is a federal tax credential earned by passing the three-part SEE, with no degree requirement. The CPA is a state licence covering accounting, audit and tax, earned through education, the CPA Exam and supervised experience. If your work is tax and only tax, the EA gets you the same standing before the IRS far faster. If you want to sign audit opinions or build an accounting career, you need the CPA.

SEE Part 3 vs Parts 1 and 2: what differs

Mon, 21 Sep 2026 00:00:00 +0000

SEE Part 3 is the part about how you practise, not how you compute. Parts 1 and 2 test tax law — income, deductions, credits, entities, returns — and reward the ability to work a number out. Part 3 tests Circular 230, preparer penalties, powers of attorney, collection and examination procedure and e-file rules, and rewards precise recall of rules, deadlines and form numbers. The format is identical across all three; the kind of thinking is not.

AAIA vs CISA: how they differ

Sun, 20 Sep 2026 00:00:00 +0000

CISA and AAIA are not alternatives. CISA — or a qualifying equivalent — is the eligibility requirement for AAIA. You hold one, then you add the other.

So the real question is what AAIA adds, and the answer is more than most CISA holders expect.

The relationship

	CISA	AAIA
Position	The information systems audit credential	Specialisation on top of it
Eligibility	Work experience in IS audit	CISA or qualifying designation
Subject	Auditing information systems	Auditing AI systems
Largest domain	Operations and resilience, and protection of information assets (26% each)	AI Operations (46%)
Questions	150, over four hours	90

CISA establishes that you can plan, execute and report an audit and form a defensible opinion. AAIA asks whether you can do that when the thing you are auditing behaves probabilistically, derives its logic from data rather than code, and can change without a deployment.

AAIR vs CRISC: how they differ

Sun, 20 Sep 2026 00:00:00 +0000

CRISC and AAIR are not competing credentials. CRISC is one of the designations that makes you eligible for AAIR, and the two are designed to stack.

The useful comparison is therefore about content: what does AAIR add to a risk professional who already holds CRISC?

The relationship

	CRISC	AAIR
Position	ISACA’s IT risk credential	AI specialisation on top
Eligibility	Work experience in IT risk	A recognised designation, CRISC included
Subject	IT risk and control	AI risk specifically
Largest domain	Risk response and reporting (32%)	AI Risk Program Management (42%)
Questions	150, over four hours	90

CRISC establishes that you can identify, assess, respond to and monitor IT risk, and that you understand controls. AAIR asks whether you can do that when the asset behaves probabilistically, derives its logic from data, changes without a deployment, and depends on a supply chain you cannot inspect.

AAISM vs CISM: how they differ

Sun, 20 Sep 2026 00:00:00 +0000

These are not alternatives, and the comparison is unusual for that reason: CISM (or CISSP) is a prerequisite for AAISM. You do not choose between them. You hold one, then you add the other.

The useful question is therefore not “which is better” but “what does AAISM add that CISM did not cover, and how much work is that?”

The relationship

	CISM	AAISM
Position	Foundational security management credential	Specialisation on top of it
Eligibility	Work experience in security management	Active CISM or CISSP required
Scope	Information security management, all of it	Security management applied to AI
Largest domain	Information security programme	AI Technologies and Controls (38%)
Questions	Longer paper	90

CISM establishes that you can govern, build and run a security programme. AAISM asks whether you can do that when the asset is a model, the supply chain includes pretrained weights and third-party datasets, and the failure modes include things software has never done before.

AB-620 vs AI-103: which agent exam?

Sun, 20 Sep 2026 00:00:00 +0000

Both exams are about building AI agents on Microsoft technology, and both sit at associate level. The difference is how you build. AB-620 is the low-code path through Copilot Studio and the Power Platform. AI-103 is the code-first path through Microsoft Foundry and Python.

Pick the one that matches the platform your organisation actually uses. They are not a sequence, and neither is a prerequisite for the other.

Side by side

	AB-620	AI-103
Certification	AI Agent Builder Associate	Azure AI Apps and Agents Developer Associate
Build surface	Copilot Studio, Power Platform	Microsoft Foundry, SDKs
Primary language	Power Fx, low-code + APIs	Python
Largest domain	Integrate and extend agents (40–45%)	Generative AI and agents (30–35%)
Distinctive content	MCP, A2A, connectors, computer use, Fabric, ALM	RAG, evaluation, vision, speech, extraction
Assumes you know	Dataverse, Power Platform, M365 Copilot	Python, Azure, REST/SDK basics
Typical role	App developer, consultant, ISV partner	Azure developer, AI engineer

What only AB-620 covers

Copilot Studio specifics: topics, agent flows, generative answers nodes, adaptive cards, variables.
Enterprise connectors: Copilot connectors, Power Platform connectors, and knowledge sources such as ServiceNow and SAP.
Computer use — agents that operate an application interface directly when no API exists.
A2A protocol for agent-to-agent collaboration, and MCP tools.
Fabric data agents.
Application lifecycle management: solutions, environment variables, Power Platform Pipelines. This is Power Platform discipline and has no equivalent on AI-103.

What only AI-103 covers

Deploying and consuming models directly, with deployment and capacity decisions.
Building RAG pipelines yourself: chunking, indexing, hybrid and vector search.
Formal evaluation — groundedness, relevance, fluency, coherence.
Computer vision, speech and information extraction as first-class domains.
Securing and monitoring the Azure infrastructure underneath.

The overlap

Real, but smaller than it looks. Both expect you to understand agents conceptually — tools, knowledge, orchestration, multi-agent patterns — and both care about responsible AI and monitoring. Both now touch Foundry: AB-620 has an entire section on integrating Copilot Studio agents with Foundry, including the model catalog and Application Insights.

Agentforce Specialist vs AB-620

Sun, 20 Sep 2026 00:00:00 +0000

These are the two major low-code agent-building certifications, and they are strikingly parallel. Agentforce Specialist builds agents on the Salesforce Platform. AB-620 builds them in Microsoft Copilot Studio.

You almost certainly do not choose between them on merit. You choose the one matching the platform your organisation runs.

Side by side

	Agentforce Specialist	AB-620
Platform	Salesforce	Microsoft Copilot Studio
Scored questions	60 (+5 unscored)	Not published
Duration	105 minutes	120 minutes
Passing score	72%	700 / 1000 (scaled)
Fee	USD 200	Set per country
Largest section	AI Agents (35%)	Integrate and extend agents (40–45%)
Renewal	Annual Trailhead module	Annual free assessment
Prerequisite	None	None

Both expect around a year of platform experience and neither enforces a prerequisite.

AI-103 vs AI-102: what changed?

Sun, 20 Sep 2026 00:00:00 +0000

AI-103 replaced AI-102 when Microsoft retired AI-102 on June 30, 2026. If you studied for AI-102, a good part of your knowledge still applies, but the focus has shifted.

The main differences

	AI-102	AI-103
Certification	Azure AI Engineer Associate	Azure AI Apps and Agents Developer Associate
Centre of gravity	Individual Azure AI services	Generative AI and agents built with Microsoft Foundry
Largest domain	Spread across services	Generative AI and agentic solutions (30–35%)
Language	C# or Python	Python

What this means for your study plan

Spend more time on agents and RAG. You should be comfortable choosing the right tools, memory and knowledge sources for an agent, and grounding answers in your own data.

AI-901 vs AI-900: what changed?

Sun, 20 Sep 2026 00:00:00 +0000

AI-901 replaced AI-900 when Microsoft retired AI-900 on June 30, 2026. Both lead to the same credential, Microsoft Certified: Azure AI Fundamentals, but the exam behind it was rebuilt rather than refreshed. If you are studying from AI-900 material, a meaningful part of it is now wrong.

The domains, side by side

AI-900 — five areas, all descriptive:

Area	Weight
Describe Artificial Intelligence workloads and considerations	15–20%
Describe fundamental principles of machine learning on Azure	15–20%
Describe features of computer vision workloads on Azure	15–20%
Describe features of Natural Language Processing workloads on Azure	15–20%
Describe features of generative AI workloads on Azure	20–25%

AI-901 — two areas, one of them hands-on:

AIF-C01 vs MLA-C02: which AWS AI exam?

Sun, 20 Sep 2026 00:00:00 +0000

These are AWS’s two AI-focused certifications below the specialty tier, and they are separated by a wide gap. AIF-C01 is foundational, for people who use AI on AWS. MLA-C02 is associate, for engineers who build and operate it.

Neither is a prerequisite for the other. Pick by what you do.

Side by side

	AIF-C01	MLA-C02
Level	Foundational	Associate
Certification	AWS Certified AI Practitioner	AWS Certified Machine Learning Engineer – Associate
Passing score	700 / 1000	720 / 1000
Scored questions	50 (of 65)	50 (of 65)
Question types	Multiple choice, multiple response, ordering, matching	Multiple choice, multiple response only
Expected experience	Up to 6 months exposure to AI/ML on AWS	1+ year with SageMaker AI and Bedrock, plus 1+ year in a related role
You are expected to	Use AI/ML solutions	Build, deploy and operate them

Note the passing scores differ. 700 for AIF-C01, 720 for MLA-C02 — a small but real difference in the bar.

AWS AI and ML certification path

Sun, 20 Sep 2026 00:00:00 +0000

Start with AIF-C01 if you are new to AI, then take MLA-C02 if you build and operate models for a living. If you already work as an ML engineer, skip AIF-C01 entirely and go straight to MLA-C02 — the foundational exam will teach you nothing you do not already use daily.

The two exams

Exam	Level	Assumes	Take it if
AIF-C01	Foundational	No hands-on ML experience	You work alongside AI projects and need to understand them
MLA-C02	Associate	Real experience building and deploying models	You are an ML or data engineer

There is no enforced prerequisite. AWS does not require AIF-C01 before MLA-C02.

CY0-001 vs Security+: which CompTIA exam do you need?

Sun, 20 Sep 2026 00:00:00 +0000

Security+ and SecAI+ are not competing exams and you do not choose between them. Security+ is CompTIA’s broad, entry-level security certification. SecAI+ (CY0-001) assumes you already have that knowledge and applies it to a single new problem: artificial intelligence, both as something to defend and as something to defend with.

If you are new to security, Security+ first. If you are already working in security and AI systems have started landing in your scope, SecAI+ is the one that teaches you something new.

Databricks Data Analyst vs Data Engineer

Sun, 20 Sep 2026 00:00:00 +0000

Databricks has two associate certifications at the data layer. The Data Analyst Associate is about querying, analysing and presenting data. The Data Engineer Associate is about getting data in and making it reliable.

Same price, same length, same validity. Different jobs.

Side by side

	Data Analyst Associate	Data Engineer Associate
Scored questions	45	45
Duration	90 minutes	90 minutes
Fee	$200	$200
Validity	2 years	2 years
Sections	9	7
Languages	English only	English, Japanese, Portuguese (BR), Korean
Largest section	Executing queries (20%)	Data Transformation and Modeling (22%)

The sections

Analyst — asking and answering questions:

Databricks Data Engineer: Associate vs Professional

Sun, 20 Sep 2026 00:00:00 +0000

Databricks offers two data engineering certifications. The Associate covers introductory data engineering on the platform; the Professional covers advanced work. Neither is a formal prerequisite for the other, though the progression is the obvious one.

Side by side

	Associate	Professional
Scored questions	45	60
Duration	90 minutes	120 minutes
Fee	$200	$200
Validity	2 years	2 years
Sections	7	9
Largest section	Data Transformation and Modeling (22%)	Developing Code using Python and SQL (23%)
Prerequisites	None	None

Same price, same validity, same format. The Professional is longer, more granular, and pitched at harder work. Its figures here follow the outline Databricks introduced on 9 October 2026.

Databricks Data Engineer: Professional vs Associate

Sun, 20 Sep 2026 00:00:00 +0000

The Associate covers introductory data engineering on Databricks. The Professional covers advanced work. The Associate is not a prerequisite — you can sit the Professional first — but the progression is the obvious one.

Side by side

	Professional	Associate
Scored questions	60	45
Duration	120 minutes	90 minutes
Sections	9	7
Fee	$200	$200
Validity	2 years	2 years
Largest section	Developing Code, Python and SQL (23%)	Data Transformation and Modeling (22%)
Prerequisites	None	None

Same price, same validity, same format. The Professional is longer and pitched harder. Professional figures are for the outline that applies from 9 October 2026; see the exam format for what changed.

Databricks GenAI Engineer vs ML Associate

Sun, 20 Sep 2026 00:00:00 +0000

Databricks has two associate-level certifications that people confuse: Generative AI Engineer Associate and Machine Learning Associate. They sit at the same level, cost the same, and test almost entirely different work.

Neither is a prerequisite for the other.

Side by side

	GenAI Engineer Associate	ML Associate
Scored questions	45	48
Duration	90 minutes	90 minutes
Fee	$200	$200
Validity	2 years	2 years
Recommended experience	6+ months with generative AI	6+ months with ML tasks
Centre of gravity	RAG applications and LLM chains	Traditional model development
Language	Python	Python (some SQL for data work)

The sections tell the story

GenAI Engineer — building an LLM application:

Databricks ML Associate vs GenAI Engineer

Sun, 20 Sep 2026 00:00:00 +0000

Databricks has two associate certifications in this space, at the same level and the same price, testing different work. ML Associate is about building models. Generative AI Engineer Associate is about building LLM applications.

Neither is a prerequisite for the other.

Side by side

	ML Associate	GenAI Engineer
Scored questions	48	45
Duration	90 minutes	90 minutes
Fee	$200	$200
Validity	2 years	2 years
Recommended experience	6+ months of ML tasks	6+ months of generative AI
Largest section	Databricks Machine Learning (38%)	Application Development (30%)
Centre of gravity	AutoML, MLflow, model development	RAG applications and chains

The sections

ML Associate — the platform and the modelling workflow:

Databricks ML: Professional vs Associate

Sun, 20 Sep 2026 00:00:00 +0000

The ML Associate tests whether you can use Databricks to perform basic machine learning tasks. The Professional tests whether you can perform advanced machine learning in production. The Associate is not a prerequisite.

Side by side

	Professional	Associate
Scored questions	59	48
Duration	120 minutes	90 minutes
Sections	3	4
Fee	$200	$200
Validity	2 years	2 years
Recommended experience	1+ years	6+ months
Languages	English only	English, Japanese, Portuguese (BR), Korean

Double the experience expectation, a longer paper, and fewer languages.

MLA-C02 vs AI-103

Sun, 20 Sep 2026 00:00:00 +0000

These two exams are both associate level and both about AI, but they test different jobs. MLA-C02 is about building, deploying and operating machine learning models. AI-103 is about building generative AI applications and agents. Choosing between them on vendor alone misses the more important difference.

Side by side

	MLA-C02	AI-103
Vendor	AWS	Microsoft
Focus	Classical ML engineering and MLOps	Generative AI apps and agents
Domains	4, near-equal weight	5, front-loaded
Heaviest content	Data prep 28%, then three domains at 24% each	Generative AI and agents 30–35%, planning 25–30%
Coding	Yes	Yes, Python
Production emphasis	Very high	High

The real difference

MLA-C02 is an operations exam wearing a machine learning badge. Its four domains carry almost equal weight, and three of them — deployment and orchestration, operating and monitoring and securing, plus data preparation — are about getting models into production and keeping them alive. Model development itself is only a quarter.

MLA-C02 vs AIF-C01: which AWS AI exam?

Sun, 20 Sep 2026 00:00:00 +0000

AWS has two AI certifications below the specialty tier, and the gap between them is wide. MLA-C02 is associate level for engineers who build and operate ML systems. AIF-C01 is foundational, for people who use AI on AWS without building it.

Neither is a prerequisite for the other.

Side by side

	MLA-C02	AIF-C01
Level	Associate	Foundational
Passing score	720 / 1000	700 / 1000
Duration	170 min (beta)	90 min
Price	75 USD (beta)	100 USD
Question types	Multiple choice, multiple response	Plus ordering and matching
Expected experience	1+ year with SageMaker AI and Bedrock, plus 1+ year in a related role	Up to 6 months exposure
You are expected to	Build, deploy and operate	Use and understand
Validity	3 years	3 years

During beta, MLA-C02 is actually cheaper than AIF-C01 despite being the harder exam. That is temporary.

N10-009 vs A+: which CompTIA exam first?

Sun, 20 Sep 2026 00:00:00 +0000

CompTIA’s recommended path is A+ → Network+ → Security+, and for Network+ it specifically recommends A+ certification plus 9 to 12 months of hands-on experience.

That is a recommendation, not a requirement. Nothing stops you booking N10-009 today.

The relationship

	A+	Network+ (N10-009)
Position	Entry point to IT	The networking step
Scope	Hardware, operating systems, software, basic networking, troubleshooting	Networking specifically, at depth
Exams required	Two	One
Typical role	Help desk, desktop support	Junior network admin, network support
Recommended before it	None	A+ plus 9–12 months experience

A+ covers a little of everything, including a small amount of networking. Network+ takes that small amount and makes it the entire exam.

SC-500 vs AZ-500: what changed?

Sun, 20 Sep 2026 00:00:00 +0000

AZ-500 and the Microsoft Certified: Azure Security Engineer Associate certification retired on August 31, 2026, along with their renewal assessments. Microsoft’s notice is blunt: “You will no longer be able to earn or renew this certification after this date.”

SC-500 is the current Microsoft exam for the security engineer role. Most of your AZ-500 knowledge still applies — but the exam was reorganised, and one substantial new area was added.

Spark Developer vs Databricks Data Engineer

Sun, 20 Sep 2026 00:00:00 +0000

Both are Databricks associate certifications at $200 for 90 minutes and 45 questions. They test different things. Spark Developer tests the engine. Data Engineer Associate tests the platform.

Side by side

	Spark Developer	Data Engineer Associate
Subject	Apache Spark itself	The Databricks platform
Largest section	DataFrame API (30%)	Data Transformation and Modeling (22%)
Architecture coverage	20%	6%
Orchestration and CI/CD	Not covered	26%
Governance	Not covered	15%
Languages	English only	English, Japanese, Portuguese (BR), Korean
Knowledge portability	High — Spark runs everywhere	Low — platform-specific

The sections

Spark Developer:

200-901 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for 200-901 if you can already read a little Python and you study 8 to 10 hours a week. The order matters: start with code, data formats and Git, move to REST APIs, which every later domain reuses, then cover Cisco platforms, and finish with deployment, automation and network fundamentals. If you have never programmed, double the timeline and spend the extra weeks on Python basics.

AB-100 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for AB-100 if you already hold an associate-level Microsoft AI, Power Platform or Dynamics 365 certification and study 8–10 hours a week. Spend week 1 on planning and strategy, week 2 on design, and weeks 3 and 4 on deployment, because Deploy is 40–45% of the exam. If Copilot Studio and Microsoft Foundry are new to you, allow six to eight weeks and build something in each.

AB-730 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks at three to five hours a week is a comfortable timeline for AB-730 if you already use Microsoft 365. Spend week 1 on generative AI fundamentals, week 2 on prompts and chats, week 3 on content and collaboration, and week 4 on agents and Cowork plus a full review. This plan follows the outline in force from October 20, 2026.

Before you start

Read Microsoft’s study guide once, end to end. The bullet lists under each objective show the level of detail the exam expects.
Get hands-on access to Microsoft 365 Copilot if you can. Many questions describe a button or a feature by name, and ten minutes in the product beats an hour of reading. If you have no licence, use the free practice assessment and the exam sandbox to see how questions are phrased.
Keep a list of look-alike pairs and add to it every session: Copilot Chat versus an agent, a prebuilt versus a custom agent, a notebook versus a page, saving a prompt versus scheduling it.

Week 1: generative AI fundamentals (25–30%)

Goal: explain how Copilot answers, and where it can go wrong.

AB-731 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

You can prepare for AB-731 in four weeks with about four to five hours a week: one week on generative AI value, one on Microsoft’s AI products, one on governance and adoption, and one on practice and gaps. The exam is conceptual and scenario-based, so the work is reading, comparing and practising decisions rather than labs. The plan assumes you are a working leader with limited time.

Week 1: generative AI and its business value (35–40%)

Goal: be able to explain, in plain words to a colleague, how generative AI works and where it pays off.

AZ-305 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for AZ-305 if you already hold AZ-104 or work with Azure and can study 8–10 hours a week. Because the exam tests design decisions, the plan swaps labs for design exercises: each week you take a short business scenario, choose services, and write down why the alternatives lose. Cover identity and governance first, data second, business continuity third, and keep infrastructure, the largest domain, for the last week when you can combine everything. Without hands-on Azure experience, add an AZ-104-style foundation first.

AZ-400 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for AZ-400 if you already build CI/CD pipelines and study 8-10 hours a week. Because build and release pipelines are 50-55% of the exam, spend two of the four weeks there: first pipelines and testing, then deployments, infrastructure as code and maintenance. Source control, work tracking, security and monitoring fit into the other two. Without pipeline experience, double the timeline and build a small project as you go.

CCNA 200-301 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for the CCNA if you already work with networks or have studied them, and can give 10 hours a week. Start with addressing and infrastructure, because everything else depends on subnetting; add switching in week 2; routing in week 3; and finish with services, security and operations in week 4. With no networking background, plan eight to twelve weeks instead.

Before you start

Read the official v2.0 exam topics once. It is three pages and lists every objective.
Install Cisco Packet Tracer or another network simulator. You will configure and break something every week.
Take any free practice test cold, such as the 20-question test on this site. The score does not matter; the weakest domain does.
Decide how you will drill subnetting. Ten minutes a day beats two hours on Sunday.

Week 1: infrastructure and connectivity (25%)

Goal: subnet in your head and read an interface.

DP-600 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

You can prepare for DP-600 in four weeks at about 8 to 10 hours a week if you already know SQL and have used Power BI. The plan below puts Prepare data first because it is almost half the exam, builds the semantic model skills in week 3, and leaves week 4 for governance, lifecycle and practice. You need a Fabric workspace to do it: a trial or a capacity, because reading alone will not work for this exam.

DP-700 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for DP-700 if you already write SQL and some PySpark and can study 8–10 hours a week. Spend the first week on the Fabric workspace, security and orchestration, the second on batch ingestion and transformation, the third on streaming and KQL, and the last on monitoring, error resolution and optimization plus practice questions. If you have never used Spark or SQL, double the timeline and start with those languages.

DP-750 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

You can prepare for DP-750 in four weeks if you already write SQL and Python and have used a Databricks workspace. The plan below spends the first week on environment and governance, the second and third on the two heavy domains (preparing data, and deploying and maintaining pipelines), and the last on tuning, troubleshooting and review. Budget about 8 to 10 hours a week and a Databricks workspace where you can build things. Without any Databricks background, double the time.

DP-800 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for DP-800 if you already write T-SQL at work and study 8–10 hours a week. Cover the material in this order: database objects and advanced T-SQL first, then security and performance, then CI/CD and APIs, then the AI topics last, because embeddings and RAG make sense only once you are comfortable with JSON, triggers and stored procedures. If you have never written T-SQL, add several weeks of SQL fundamentals first.

How to study for an IT certification exam

Fri, 09 Oct 2026 00:00:00 +0000

The most reliable way to prepare for an IT certification exam is to work from the official skills outline, learn each topic by doing it, test yourself often with realistic questions, and spend the final weeks on your weakest domain. Reading or watching alone gives you the feeling of progress without the recall you need on the day. This page lays out a method that works for any exam on this site.

IT certifications with no IT experience

Fri, 09 Oct 2026 00:00:00 +0000

With no IT experience, the exams on this site that assume least are Microsoft’s AZ-900 and SC-900 and AWS’s AIF-C01. None enforces a prerequisite, none requires coding for a pass, and each can be prepared in about four weeks at a few hours a week. The exams to leave until later are the ones whose own pages assume experience, such as Security+, the CCNA and every associate-level exam.

PL-300 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for PL-300 if you already open Power BI Desktop at work and can study 8–10 hours a week. The plan follows the exam’s own loop: prepare the data in week 1, model it in week 2, visualise it in week 3, and manage and secure it in week 4, with a practice-test review at the end. Without any experience of Power BI, double the timeline, and spend the extra time on weeks 1 and 2.

Port numbers flashcards and drill

Fri, 09 Oct 2026 00:00:00 +0000

Port numbers are memory work, so the fastest way to learn them is retrieval, not rereading. Choose a mode, answer, and the drill tells you at once whether you were right. The reference list under the drill groups the ports by purpose.

Reference list

Patterns that make them easier

Most secure variants sit next to or near the plain port: 80 and 443, 389 and 636, 110 and 995, 143 and 993.
SSH, SFTP and SCP all share port 22; FTP uses 21 for control and 20 for data in active mode.
Mail has three jobs: SMTP sends (25 between servers, 587 or 465 for clients), POP3 and IMAP fetch.

The annotated list is in the common ports cheat sheet. Ports are tested in Network+ N10-009 and Security+ SY0-801; try the port practice questions for scenario wording.

SC-200 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for SC-200 if you already work with Microsoft 365 security or Azure and study 8–10 hours a week. Start with KQL, because it runs through all three domains; then build the Sentinel and Defender XDR environment; then practise detections and automation; and finish with incident response and review. Without any security operations experience, double the timeline and consider SC-900 first.

Before you start

Read the official skills outline once, end to end. It lists every objective and is the document the exam is written against.
Set up a lab. A Microsoft 365 trial tenant and an Azure subscription are enough. Create a Log Analytics workspace, onboard Microsoft Sentinel to it, and set a budget alert so ingestion does not surprise you.
Take the free Microsoft practice assessment cold. The score does not matter yet; the per-domain breakdown does.
Bookmark the KQL reference. You will use it every day for four weeks.

Week 1: KQL and hunting (20–25%, but it underpins everything)

Goal: read and write queries without looking up the syntax.

SC-300 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for SC-300 if you already administer Microsoft 365 or Entra ID and study 8–10 hours a week. Follow the order of the outline: users and hybrid identity first, then authentication and Conditional Access (the largest domain), then applications and workload identities, then governance and monitoring. Without any Microsoft 365 administration experience, double the timeline and look at the SC-900 guide first.

Before you start

Read the official skills outline once, end to end. Note the version date: the English outline changes on October 28, 2026.
Get a test tenant. A Microsoft 365 developer or trial tenant with Entra ID P2 and Entra ID Governance features turned on lets you try PIM, ID Protection and entitlement management. Licensing requirements vary, so check which licence each feature needs before you build a lab.
Take the free Microsoft practice assessment cold. The score does not matter yet; the per-domain breakdown does.
Create two cloud-only emergency access accounts early. Creating them is an exam objective.

Week 1: user identities (20–25%)

Goal: build the tenant, the identities and the hybrid link.

SC-401 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for SC-401 if you already administer Microsoft 365 and can study 8–10 hours a week. Spend week one on classification and sensitivity labels, week two on DLP and Endpoint DLP, week three on retention and Insider Risk Management, and week four on audit, eDiscovery, DSPM for AI and review. If you have never opened the Microsoft Purview portal, add two weeks of hands-on time first, because this exam rewards people who have clicked through the settings.

SC-900 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for SC-900 at 4–6 hours a week, even if you start without a security background. Cover the four skill areas in an order that builds on itself: concepts and Microsoft Entra first because every later product assumes them, then the largest area (security solutions), then compliance and review. Weights decide where the hours go: security solutions is 35–40%, so it gets the most.

Before you start

Read the official skills outline once. It is short and lists every objective.
Take the free Microsoft practice assessment cold. The score does not matter; the per-area breakdown does.
Use a free Microsoft 365 or Azure trial tenant if you can. Looking at the Entra admin center, the Defender portal and the Purview portal makes product names stick. It is optional.

Week 1: concepts and Microsoft Entra (concepts 10–15%, Entra 25–30%)

Goal: speak the vocabulary and understand how identity is protected.

SY0-801 study plan: pass in 4 weeks

Fri, 09 Oct 2026 00:00:00 +0000

Four weeks is realistic for SY0-801 if you already work in IT and study 8-10 hours a week. Start with general concepts and threats because every later domain uses that vocabulary, then architecture, then Security Operations (the largest domain), and finish with governance and a review of AI topics. If you have no IT background, double the timeline and consider Network+ N10-009 first.

Before you start

Download the official SY0-801 exam objectives from CompTIA and read them once. They list every topic, plus an acronym list you should know.
Set up a lab you can break: two virtual machines (one Windows, one Linux), a free SIEM or log viewer, and Wireshark.
Take a diagnostic set of questions cold. The score does not matter; the weak domains do.
Check whether SY0-801 is available in your region yet. It is expected around November 17, 2026.

Week 1: concepts, threats and attacks (16% + 24%)

Goal: name any control, actor, attack or indicator in a scenario.

AZ-802 study plan: pass in 4 weeks

Tue, 29 Sep 2026 00:00:00 +0000

Four weeks is realistic for AZ-802 if you already administer Windows Server and study 8–10 hours a week. Cover the seven domains in an order that builds a lab as you go: AD DS first, because every later exercise needs a domain; then remote management, Azure Arc and virtual machines; then DNS, DHCP and storage; and finally security and troubleshooting, which reuse everything before them. Without Windows Server experience, plan on eight weeks or more.

AB-900 study plan: pass in 4 weeks

Mon, 28 Sep 2026 00:00:00 +0000

Four weeks at five to six hours a week is realistic for AB-900 if you already use Microsoft 365 at work and have seen at least one admin center. Spend the first week on the tenant itself, the second on data protection, the third on Copilot and agents, and the fourth on review. If you have never administered Microsoft 365, allow six weeks and spend the extra time in weeks one and two.

AI-300 study plan: pass in 4 weeks

Mon, 28 Sep 2026 00:00:00 +0000

Four weeks is realistic for AI-300 if you already train models in Python, have used Azure Machine Learning, and can study 8–10 hours a week. Build the infrastructure first, then run the machine learning lifecycle on it, then do the same for generative AI in Foundry, and finish with evaluation and optimisation. Without Azure Machine Learning experience, or without any GitHub Actions background, plan on six to eight weeks.

Before you start

Read the official skills outline end to end. It lists every objective in five short sections.
Create an Azure subscription with a budget alert. Compute clusters scale to zero, but compute instances, managed online endpoints and provisioned deployments bill while they exist. Delete resource groups at the end of each session.
Install the Azure CLI with the ml extension, Python with the azure-ai-ml and mlflow packages, and create a GitHub repository you can push workflows to.

Week 1: MLOps infrastructure (15–20%)

Goal: a workspace you can recreate from code, reached securely from GitHub.

AZ-104 study plan: pass in 4 weeks

Mon, 28 Sep 2026 00:00:00 +0000

Four weeks is realistic for AZ-104 if you already work with Azure or another cloud and study 8–10 hours a week. Cover the five domains in an order that builds on itself: identity and governance first, because every later lab needs roles and resource groups; then storage; then compute, the largest block; then networking and monitoring together in the last week. Without any cloud or server administration experience, double the timeline.

AZ-900 study plan: pass in 4 weeks

Mon, 28 Sep 2026 00:00:00 +0000

Four weeks is a comfortable timeline for AZ-900 at four to six hours a week, even with no cloud experience. Spend week 1 on cloud concepts, weeks 2 and 3 on Azure architecture and services because it is the largest domain, and week 4 on management and governance plus review. If you already use Azure at work, two weeks is usually enough.

Before you start

Read the official skills outline once, end to end. It is two pages and every line starts with “describe”, which tells you the depth expected.
Create a free Azure account if you can. You do not need to build anything for AZ-900, but clicking through the portal makes the names stick. Set a budget alert on day one and delete what you create.
Keep a two-column list of commonly confused pairs: availability set versus availability zone, VPN Gateway versus ExpressRoute, Azure Policy versus RBAC, and so on. Add to it every session. It becomes your revision sheet.

Week 1: cloud concepts (25–30%)

Goal: explain the cloud to someone who has never used it.

AI-200 study plan: pass in 4 weeks

Sun, 27 Sep 2026 00:00:00 +0000

Four weeks is realistic for AI-200 if you already write Python against Azure and study 8–10 hours a week. Spend one week per domain, in an order that builds on itself: containers first so you have somewhere to run code, then the data services it talks to, then the messaging that connects the pieces, then security and monitoring across all of it. Without Azure development experience, double the timeline.

Before you start

Read the official skills outline once, end to end. It is short.
Create a free or pay-as-you-go Azure subscription and set a budget alert. AKS, PostgreSQL and Managed Redis cost money when left running; delete resource groups at the end of each session.
Set up Python with the Azure SDKs and the Azure CLI. Almost every lab below uses DefaultAzureCredential, so sign in with az login once and let it pick up your identity.

Week 1: containerized solutions (20–25%)

Goal: build an image, store it, and run it three ways.

SEE Part 1 study plan: pass in 8 weeks

Mon, 21 Sep 2026 00:00:00 +0000

Plan on eight weeks at 8–10 hours a week for SEE Part 1. That is honest for someone who knows basic tax but does not prepare returns for a living. If you prepare individual returns every season, five or six weeks is realistic; if you are starting from zero, allow ten to twelve. Part 1 covers six domains from dependency tests to estate tax, and four weeks is rarely enough to learn that breadth and still practise under exam conditions.

SEE Part 2 study plan: pass in 6 weeks

Mon, 21 Sep 2026 00:00:00 +0000

Six weeks at 8–10 hours a week is a realistic plan for SEE Part 2 if you already prepare returns or have passed Part 1. If business returns are new to you — no Schedule C, 1065 or 1120-S experience — stretch it to eight weeks by giving weeks 1 and 3 a second week each. Four weeks is not honest for this part: the entity rules are dense, the basis calculations need practice rather than reading, and 18 questions of specialized returns cover five unrelated topics.

SEE Part 3 study plan: pass in 6 weeks

Mon, 21 Sep 2026 00:00:00 +0000

Six weeks at 8–10 hours a week is an honest target for SEE Part 3 if you have not worked in representation before. The content is less computational than Parts 1 and 2 but more unfamiliar: most preparers have never handled a collection case, an appeal or an e-file infraction. If you already represent clients before the IRS, four weeks is realistic; compress weeks 3 and 4.

The plan follows the four domains in order of weight, then leaves a full week for practice, because Part 3 is won by recall under the EXCEPT format.

AAIA study plan: pass in 6 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 6–8 hours a week, assuming you hold CISA or a qualifying designation — you must, to certify.

The plan deliberately inverts most auditors’ instincts. AI Auditing Tools and Techniques is only 21% of the exam and you already have that discipline. AI Operations is 46% and is mostly new. Study time follows the weighting, not the comfort.

Week 1: AI governance and risk (33%), part one

AI models, considerations and requirements. What a model is, the difference between training and inference, why behaviour comes from data. You cannot audit what you cannot describe.
AI governance and programme management. Inventories, ownership, approval paths, oversight structures.

Week 2: governance and risk, part two

AI risk management. Risk sources specific to AI, thresholds expressed as measurements rather than pass/fail, treatment decisions, and why reassessment is mandatory when drift exists.
Privacy and data governance programmes. Training data as personal data, minimisation, retention, and the awkward fact that deleting a record does not remove its influence from a trained model.
Leading practices, ethics, regulations and standards. Not clause recitation — the consistent shape: inventory, impact assessment, human oversight, documentation, transparency, auditability.

Weeks 3, 4 and 5: AI operations (46%)

Three weeks for nearly half the exam.

AAIR study plan: pass in 6 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 6–8 hours a week, assuming you hold a qualifying designation and already work in risk.

The allocation follows the weighting, which is not where most candidates’ anxiety points. The lifecycle domain sounds the most technical and is the smallest at 21%. Programme management is 42% and governance 37% — nearly four fifths of the exam is risk discipline applied to a new asset.

Weeks 1 and 2: AI risk governance and framework integration (37%)

Week 1 — models, processes, accountability:

AAISM study plan: pass in 6 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 6–8 hours a week. This plan assumes you hold CISM or CISSP — you must, to certify — and therefore that security management methodology is already yours. What you are learning is how it changes when the asset is a model.

The plan is deliberately weighted toward the technologies domain. It is 38% of the exam, the largest of the three, and the one experienced managers consistently underprepare.

AB-620 study plan: pass in 4 weeks

Sun, 20 Sep 2026 00:00:00 +0000

This plan assumes you already build agents in Copilot Studio at a basic level and can give it 8–10 hours a week. If you have never configured an agent with knowledge sources, instructions, tools and topics, start with the prerequisites — Microsoft assumes that skill before the exam begins.

You need a Copilot Studio environment you can build in. This exam cannot be passed from reading alone; the integration domain is 40–45% and it is all mechanisms you have to have wired up at least once.

Agentforce Specialist study plan

Sun, 20 Sep 2026 00:00:00 +0000

Five weeks at 6–8 hours a week, assuming roughly the year of Salesforce platform experience Salesforce expects. You need an org you can build in — a Developer Edition or a sandbox with Agentforce enabled.

The plan follows the weighting: AI Agents, Prompt Engineering and Data 360 are 75% of the exam between them.

Week 1: prompt engineering (20%)

When Prompt Builder is appropriate given a business requirement — and when an agent action or a flow is the better answer.
Prompt template types, including field generation and flex templates, and the considerations for each.
Creating, activating and executing prompt templates. Activation is a distinct step and it is examinable.
Access controls governing prompt templates — who can see, edit and use them.
Writing effective prompts: clarity, structure, examples, constraints.
Grounding techniques — choosing the right one for a scenario.
The Trust Layer: its security and privacy features, and what it does to data on the way to and from a model.
Managing and preventing access to specific models.

Week 2: Data 360 fundamentals (20%)

The section most candidates underestimate, because it is conceptual rather than clicky.

AI-103 study plan: pass in 4 weeks

Sun, 20 Sep 2026 00:00:00 +0000

This plan assumes you already write Python and have basic Azure experience, and that you can study about 8–10 hours a week. Short on experience? Stretch each week to two.

Week 1: plan and manage an Azure AI solution

Set up a Microsoft Foundry project and deploy a model.
Secure it with managed identity and restrict network access.
Turn on monitoring and look at token usage, latency and traces.
Learn how to choose between models and deployment types based on cost and requirements.

Week 2: generative AI and agents

Build an agent with at least two tools, such as a search tool and a custom function.
Add RAG: index your own documents and ground the agent’s answers in them.
Run an evaluation on relevance, groundedness and safety.
Add content filters and see what they block.

Week 3: vision, text and information extraction

Analyse images and add image understanding to an agent.
Build a text pipeline: language detection, entities, sentiment, translation.
Extract fields from PDFs and forms into structured output.

Week 4: practice and fix gaps

Take a full practice exam under real conditions: 120 minutes, no notes.
Review every wrong answer and note the domain it belongs to.
Revisit the weakest domain, then take a second practice exam.
Book the exam once you consistently score above 80%.

AI-901 study plan: pass in 4 weeks

Sun, 20 Sep 2026 00:00:00 +0000

This plan assumes about 5–7 hours a week and some exposure to Azure. If you have never written a line of Python, add a week at the front for basic syntax — the exam assumes it.

The weighting drives the plan. Concepts are 40–45% of the score, implementation 55–60%. Most people invert this, spend three weeks reading about responsible AI, and then meet a question about the Foundry SDK they have never opened.

AIF-C01 study plan: pass in 4 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Four weeks at 5–7 hours a week. This plan assumes some familiarity with AWS but no AI background, and it follows the weighting rather than the domain order — generative AI and foundation models are 52% of the scored content.

You need an AWS account you can experiment in. Bedrock and SageMaker AI appear throughout and are far easier to learn by opening than by reading about.

Week 1: fundamentals of AI and ML (20%)

The vocabulary everything else rests on.

CY0-001 study plan: pass CompTIA SecAI+ in 4 weeks

Sun, 20 Sep 2026 00:00:00 +0000

This plan assumes you already work in security and can give it 6–8 hours a week. If AI is entirely new to you, add a week at the front for concepts. If security is new to you, this is the wrong exam to start with — see CY0-001 vs Security+.

The plan follows the domain weights rather than the order they appear in the outline, because one domain is worth 40% and the others are not.

Databricks certification path

Sun, 20 Sep 2026 00:00:00 +0000

Start with the associate exam for the job you actually do — Data Analyst, Data Engineer, Machine Learning or Spark Developer — and only then consider a professional exam. The professional exams are a genuine step up, not a slightly harder version of the associate.

The seven exams

Exam	Level	Who it is for
Data Analyst Associate	Associate	Analysts querying and visualising with Databricks SQL
Data Engineer Associate	Associate	Engineers building ingestion and transformation pipelines
Data Engineer Professional	Professional	Engineers running production pipelines at scale
Machine Learning Associate	Associate	Practitioners training and tracking models on the platform
Machine Learning Professional	Professional	Practitioners running models in production
Generative AI Engineer Associate	Associate	Engineers building RAG and agent applications
Spark Developer Associate	Associate	Developers writing Spark DataFrame code

Pick the lane, not the ladder

These are not seven rungs. They are three or four separate lanes, each with an associate entry point and, in two cases, a professional exam above it.

Databricks Data Analyst Associate study plan

Sun, 20 Sep 2026 00:00:00 +0000

Five weeks at 6–8 hours a week. You need a Databricks workspace with Databricks SQL access and some data to query — nine sections is a lot of surface, and most of it is hands-on.

The plan groups the nine sections rather than treating each individually.

Week 1: the platform (11%), managing and importing data (13%)

The foundational sections, 24% together.

The Databricks Data + AI Platform: the lakehouse idea, Delta tables, how the workspace is organised, and where Databricks SQL sits within it.
SQL warehouses: what they are, how they differ from all-purpose compute, sizing, and the fact that they cost money while running.
Managing data: catalogs, schemas and tables; creating and altering them; managed versus external tables.
Importing data: loading files and connecting sources. Only about two questions — cover it and move on.

Week 2: executing queries (20%)

The largest section.

Databricks Data Engineer Associate study plan

Sun, 20 Sep 2026 00:00:00 +0000

Five weeks at 6–8 hours a week. You need a Databricks workspace — this is a platform exam and the questions assume you have used Delta Lake, Lakeflow Jobs and Unity Catalog rather than read about them.

Because no section exceeds 22%, the plan is about coverage. Seven sections, none of which you can safely skip.

Week 1: the platform (6%) and ingestion (21%)

The platform section is only about three questions — cover it quickly:

Databricks Data Engineer Professional study plan

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 8–10 hours a week. Nine sections is a lot of surface, so this plan groups them rather than treating each separately. It follows the outline that applies from 9 October 2026; the exam format page lists what changed.

You need a Databricks workspace with enough data to make performance visible. Several sections — optimisation, debugging, monitoring — cannot be learned on toy datasets where everything is instant.

Hold the Associate already? See the compressed plan at the end.

Databricks GenAI Engineer study plan

Sun, 20 Sep 2026 00:00:00 +0000

Five weeks at 6–8 hours a week, assuming Databricks recommends 6+ months of hands-on generative AI experience and you have roughly that. You need a Databricks workspace you can build in — this exam rewards having shipped something.

The plan follows the weighting. Application development and deployment are 52% between them; governance is 8%.

Week 1: design applications (14%) and data preparation (14%)

The two smaller front-end sections together.

Design:

Databricks ML Associate study plan

Sun, 20 Sep 2026 00:00:00 +0000

Five weeks at 6–8 hours a week, assuming roughly the 6+ months of hands-on experience Databricks recommends. You need a Databricks workspace — the largest section is the platform itself, and it cannot be learned from reading.

The plan follows the weighting. Databricks Machine Learning is 38%, Model Development 31%, ML Workflows 19%, Model Deployment 12%.

Weeks 1 and 2: Databricks Machine Learning (38%)

The largest section, and the one people wrongly treat as background. Two weeks.

Databricks ML Professional study plan

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 8–10 hours a week, assuming roughly the one year of hands-on experience Databricks recommends. You need a workspace and, ideally, a model you have run long enough to have seen behave badly.

The plan follows the 44/44/12 weighting: half the time on model development, half on ML Ops, and a short block on deployment.

Weeks 1 and 2: model development (44%), part one and two

Week 1 — rigorous development:

Databricks vs Azure data engineering

Sun, 20 Sep 2026 00:00:00 +0000

Certify where you work. If your pipelines run on Databricks, take the Data Engineer Associate; if your organisation is Azure-first without Databricks, a Microsoft credential is the better signal. The two are not interchangeable, and a certification in a platform you do not touch is worth very little in an interview.

What each one signals

	Databricks	Azure
Certifies	Platform-specific pipeline engineering on the lakehouse	Broader cloud data platform work
Core concepts	Delta Lake, medallion architecture, Lakeflow jobs, Unity-style governance	Azure data services, integration, security posture
Spark knowledge	Assumed and tested	Less central
Professional tier	Yes, a genuine step up	No direct equivalent; Microsoft uses associate and expert levels
Passing score published	No	Yes, typically 700/1000

The overlap is real but shallow

Both worlds share the same underlying ideas: incremental ingestion, idempotent writes, schema evolution, partitioning and file layout, lineage, quality gates, and the discomfort of a pipeline that used to take twenty minutes and now takes two hours.

MLA-C02 study plan: pass in 6 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 8–10 hours a week, assuming you meet AWS’s expectation of roughly a year with SageMaker AI and Bedrock. If you do not, see the prerequisites — this is not an exam you can read your way through.

You need an AWS account you can build in. Every domain here is about doing something, and the questions are written by people who assume you have.

Weeks 1 and 2: data preparation for ML and AI (28%)

The largest domain, and the one engineers coming from a modelling background underrate.

N10-009 study plan: pass Network+ in 6 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks at 6–8 hours a week. This assumes some IT exposure — CompTIA recommends A+ plus 9 to 12 months in a support role, and the plan is realistic for someone around that level.

Two things run through every week: subnetting practice and performance-based question practice. Both are skills rather than knowledge, and both need repetition rather than reading.

Week 1: networking concepts, part one (23%)

The OSI model — seven layers, what lives at each, and which device or protocol belongs where. This is the mental framework the whole exam uses.
The TCP/IP model and how it maps across.
Topologies and network types: LAN, WAN, WLAN, and the cloud models.
Cables and connectors, and the distance and speed limits that make a scenario answerable.

Week 2: networking concepts, part two, and subnetting

IPv4 addressing: classes, private ranges, CIDR notation.
Subnetting — the skill most people underestimate. Practise until you can subnet without a calculator, quickly.
IPv6 fundamentals.
Ports and protocols: the common ones, their numbers, and what they do. These recur across every other domain.
DHCP, DNS and NAT.

Subnetting deserves daily practice from here to the exam. Ten minutes a day beats one long session.

SC-500 study plan: pass in 6 weeks

Sun, 20 Sep 2026 00:00:00 +0000

Six weeks rather than four, because SC-500 is wide and its domains are evenly weighted. There is no dominant area to prioritise and no small area to skip — every domain is worth roughly a quarter of the score. This plan assumes 8–10 hours a week and existing Azure administration experience.

Held AZ-500? Skip to the compressed plan at the bottom.

Week 1: identity, access and governance (20–25%)

Entra ID: PIM, conditional access, MFA and passwordless, enterprise applications and app registrations, OAuth permission grants and consent settings, managed identities.
Key Vault: deploy it, configure access and firewall settings, manage keys, secrets and certificates, scan for secrets with Defender CSPM, enable Defender for Key Vault.

Week 2: governance

Still domain one, but it deserves its own week — it is the part experienced engineers most often assume they know.

Spark Developer Associate study plan

Sun, 20 Sep 2026 00:00:00 +0000

Five weeks at 6–8 hours a week. You need an environment where you can run PySpark against data large enough that performance is visible — architecture and tuning questions make little sense on datasets that fit in memory.

Half the exam is writing transformations, so half the plan is too.

Week 1: architecture and components (20%)

Deliberately first, because everything else depends on it.

The execution model: driver and executors, and what each does.
Jobs, stages and tasks — and what causes a new stage to begin.
Partitions — the unit of parallelism, and why partition count matters.
Narrow versus wide transformations. Narrow transformations operate within a partition; wide ones require a shuffle, moving data across the cluster. This is the single most important idea in the exam.
Lazy evaluation — transformations build a plan, actions trigger execution.
The Catalyst optimiser and adaptive query execution at a conceptual level.
Caching and persistence: what they do and what they cost.

Weeks 2 and 3: the DataFrame API (30%)

The largest section, two weeks.

Binary and hex conversion drill

Fri, 09 Oct 2026 00:00:00 +0000

Fast binary conversion is the base skill under subnetting and IPv6. This drill covers 8-bit decimal and binary, two-digit hexadecimal, and the mask octets 128 to 255.

The method

Binary place values for one octet, left to right: 128, 64, 32, 16, 8, 4, 2, 1. Add the values that have a 1.
Hexadecimal digit = 4 bits. A two-digit hex number is high digit × 16 + low digit; for example C0 is 12 × 16 + 0 = 192.
Mask octets only come in eight forms: 128, 192, 224, 240, 248, 252, 254, 255, adding one leading 1-bit at a time.

Use the results in subnetting step by step and in the subnetting trainer. Binary is tested in CCNA 200-301 and Network+ N10-009.

DP-600 sample questions: 5 free questions

Fri, 09 Oct 2026 00:00:00 +0000

Five original scenario questions in the style of the DP-600 skills outline: governance, choosing a data store, T-SQL, Direct Lake and DAX. Read the requirement in each scenario first, then eliminate options that break it. Click “Show answer” after you commit. There is a longer set in the free practice test.

Maintain a data analytics solution

Question 1. A group of finance analysts must query the curated tables in one lakehouse through its SQL analytics endpoint. They must not edit anything and must not see the other items in the workspace. What should you do?

Free 200-901 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free 200-901 sample questions drawn from five of the six domains, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official exam topics, not questions from the real exam. The sixth domain, software development and design, is covered in the practice test.

Question 1. A Python script sends a POST request to create a resource. The API replies with HTTP 400 and a JSON body saying that a required field is missing. What should the developer do?

Free AB-100 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free AB-100 sample questions, covering the Plan, Design and Deploy domains, with the answer and the reasoning. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam. Deploy gets three, because it is 40–45% of the exam.

Question 1. A retailer wants an assistant for store managers. A prebuilt agent covers about 80% of the requirement. The remaining 20% is a lookup in an internal stock system. The business wants the shortest time to value and the lowest running cost. What should you recommend?

Free AB-730 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free AB-730 sample questions covering all four skill areas, with the answer and the reasoning behind it. Try each one before reading the explanation. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A sales manager asks Copilot to summarise a market report and gets a confident paragraph that includes a statistic she cannot find in the report. What is the most appropriate next step?

Free AB-731 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free AB-731 sample questions covering all three skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Identify the business value of generative AI solutions

Question 1. A company pilots a generative AI assistant that summarises long contracts. The monthly bill is far above the forecast, even though the number of users is as planned. Which factor most directly explains the usage cost?

Free AZ-305 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free AZ-305 sample questions, one or more from each design area, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A company has 20 subscriptions. The security team needs to query logs from all of them in one place, keep them for one year, and run cross-resource queries with KQL. Which design should you recommend?

Free AZ-400 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free AZ-400 sample questions, one from each skill area, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A team uses Azure Boards for planning and GitHub for code. Auditors want every production change traceable to a requirement, without developers keeping a spreadsheet. What should you configure?

Free CCNA 200-301 sample questions

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free CCNA 200-301 sample questions, one per major area of the v2.0 blueprint, with the answer and the reasoning behind it. Try each question before opening the answer. These are original scenario questions written against the published exam topics, not questions from the real exam.

Question 1. A host is configured with the address 172.16.45.200/20. What is the broadcast address of its subnet?

A. 172.16.32.255
B. 172.16.47.255
C. 172.16.45.255
D. 172.16.48.0

Show answer

Answer: B

Free DP-700 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free DP-700 sample questions with the answer and the reasoning behind it. Try each one before opening the answer. They are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A workspace for development, one for test and one for production are connected through a deployment pipeline. Notebooks in each stage must use the lakehouse that belongs to that stage, not the development lakehouse. What should you configure?

Free DP-750 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free DP-750 sample questions covering all four skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A transformation notebook runs once every night for about 40 minutes, with no interactive users. The team wants to pay for compute only while the run is active and not manage a long-lived cluster. Which compute choice fits best?

Free DP-800 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free DP-800 sample questions covering the three skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. An Orders table stores each order's lines as a JSON array in a column named Lines, for example [{'sku':'A1','qty':2},{'sku':'B7','qty':1}]. A report needs one row per line, with sku as a string and qty as an integer. Which approach produces that result?

Free PL-300 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free PL-300 sample questions, one or more from each skill area, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. You have two Power Query queries, Sales2024 and Sales2025. They have identical columns and each holds one year of orders. You need a single table that contains every order from both years. What should you use?

Free SC-200 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free SC-200 sample questions, covering the three skill areas, with the answer and the reasoning behind each. Try each one before opening the answer. These are original SOC scenario questions written against the official skills outline, not questions from the real exam.

Question 1. Your SOC wants every new Microsoft Sentinel incident with severity High to be assigned to the on-call analyst group and tagged, without any custom code. Another step must also send a message to a chat channel using a connector. How should you implement this?

Free SC-300 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free SC-300 sample questions covering all four skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A regional IT team must reset passwords for users who belong to the Sales department only. They must not be able to reset passwords for any other users in the tenant. What is the best way to scope their access?

Free SC-401 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free SC-401 sample questions, one from each major area of the skills outline, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. Contoso wants every document in a SharePoint library that contains a customer-number pattern to be labelled Confidential automatically, including the millions of files already stored there. Users must not be prompted. What should you configure?

Free SC-900 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free SC-900 sample questions, covering all four skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A company decides that no user, device or network location is trusted by default, and that every request is verified explicitly using all available signals. Which security model does this describe?

Free SY0-801 sample questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Here are five free SY0-801 sample questions, one from each domain, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official exam objectives, not questions from the real exam.

Question 1. A company puts signs on the server room door stating that all entries are recorded and that unauthorized access will be prosecuted. Which control type best describes the signs?

Help desk to networking or security certification

Fri, 09 Oct 2026 00:00:00 +0000

A help desk technician should take networking next, unless security duties are already part of the job. Network+ N10-009 sits between A+ and Security+ in CompTIA’s core sequence, and the site’s CompTIA path says to take it before any security certification. Security+ SY0-801 is the better next exam only if you already handle accounts, logs, patching or incidents and have a networking base.

Useful links for certification exam prep

Fri, 09 Oct 2026 00:00:00 +0000

These are the official sources worth bookmarking when you prepare for an IT or tax certification exam: the certification bodies’ own exam pages and skills outlines, the testing providers, and a few free tools. Always check the official page for the current price, version and rules, because those change. We are not affiliated with any of the organisations below.

Certification bodies

Organisation	What you find there
Microsoft Learn: credentials	Exam pages, study guides with the skills measured, free learning paths, renewal
Microsoft certifications	Every Microsoft certification and exam, with prerequisites and languages
CompTIA certifications	Network+, Security+, SecAI+ and others, with exam objectives
Cisco Learning Network	CCNA and related exam topics, forums and study material
Cisco DevNet	Sandboxes and learning labs for network automation
AWS Certification	AWS exam guides, practice material and exam policies
Databricks certification	Databricks exam guides and registration
ISACA credentialing	AAIA, AAIR, AAISM and the other ISACA credentials
Salesforce Trailhead credentials	Salesforce certifications and learning trails
IRS: enrolled agents	How to become an enrolled agent and the Special Enrollment Examination

Practise the exam interface

Microsoft exam sandbox lets you try the question types and the exam interface before the day.
Microsoft practice assessments are free official question sets for many Microsoft exams.
Microsoft exam scoring and score reports explains what a scaled score of 700 means.
Request exam accommodations at Microsoft covers extra time and assistive devices.
Microsoft certification renewal explains the free annual assessment.

Booking the exam

Pearson VUE delivers many of the exams covered on this site, at test centres and online. Which provider runs your exam is stated on the certification body’s exam page. Our exam day guide covers identification, the room check and retakes.

Free AZ-802 sample questions with answers

Tue, 29 Sep 2026 00:00:00 +0000

Here are five free AZ-802 sample questions from five of the seven skill areas, each with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A branch office has a read-only domain controller. Branch staff must be able to sign in when the WAN link is down, but the credentials of domain administrators must never be stored on the RODC. What should you configure?

Free AB-900 sample questions with answers

Mon, 28 Sep 2026 00:00:00 +0000

Here are five free AB-900 sample questions covering all three skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A new marketing team of 25 people needs Microsoft 365 apps. People join and leave the team every month, and the admin wants licences to follow team membership without manual work. What should the admin do?

Free AI-300 sample questions with answers

Mon, 28 Sep 2026 00:00:00 +0000

Here are five free AI-300 sample questions, one from each skill area, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A GitHub Actions workflow deploys Azure Machine Learning resources with Bicep. The security team forbids storing any long-lived Azure credential in GitHub. How should the workflow authenticate to Azure?

Free AZ-104 sample questions with answers

Mon, 28 Sep 2026 00:00:00 +0000

Here are five free AZ-104 sample questions, one from each skill area, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A developer must be able to create and delete any resource in the resource group RG-Dev, but must not be able to grant access to anyone else or touch any other resource group in the subscription. Which assignment meets the requirement with the least privilege?

Free AZ-900 sample questions with answers

Mon, 28 Sep 2026 00:00:00 +0000

Here are five free AZ-900 sample questions covering all three skill areas, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A company moves its customer database from its own servers to a platform-as-a-service database on Azure. Which responsibility stays with the company in every cloud service type?

Free AI-200 sample questions with answers

Sun, 27 Sep 2026 00:00:00 +0000

Here are five free AI-200 sample questions, one or more from each skill area, with the answer and the reasoning behind it. Try each one before opening the answer. These are original scenario questions written against the official skills outline, not questions from the real exam.

Question 1. A Python API runs in Azure Container Apps. It must scale out when messages pile up in an Azure Service Bus queue and scale to zero replicas when the queue is empty. What should you configure?

Free SEE Part 1 sample questions

Mon, 21 Sep 2026 00:00:00 +0000

Here are five free SEE Part 1 questions with explained answers, one each from five of the six domains. They are original, scenario-based questions written at the level of the official content outline, not real exam items. Like the exam, they assume tax year 2025 unless stated otherwise, and each has four options and one best answer.

Try each question before you open the answer. If you can explain why the three wrong options are wrong, you know the rule well enough for the exam.

Free SEE Part 2 sample questions with answers

Mon, 21 Sep 2026 00:00:00 +0000

Here are five free SEE Part 2 practice questions covering all three domains, each with a worked explanation. They are original questions written to the level of the published content outline, not real exam items, and they follow the three formats the exam uses: direct question, incomplete sentence and “all of the following EXCEPT”. All figures relate to tax year 2025, as the current exam does.

Question 1. Alex contributes land to a newly formed partnership in exchange for a 25% interest. The land has an adjusted basis to Alex of $40,000 and a fair market value of $70,000, and is not subject to any liability. What is Alex's basis in the partnership interest immediately after the contribution?

Free SEE Part 3 sample questions

Mon, 21 Sep 2026 00:00:00 +0000

These five questions cover all four SEE Part 3 domains, in the formats PSI uses: direct questions, incomplete sentences and “all of the following EXCEPT”. They are original scenarios written to the content outline, not real exam items. Each answer explains the rule, because in Part 3 the rule is the whole point.

Question 1. An enrolled agent prepared a client's return and later discovers the client omitted a Form 1099-NEC from it. Under Circular 230, what must the enrolled agent do?

AB-620 vs Agentforce Specialist

Sun, 20 Sep 2026 00:00:00 +0000

Both exams certify that you can build AI agents without writing much code, and both are decided entirely by which platform your organisation uses. AB-620 is Microsoft Copilot Studio; Agentforce Specialist is Salesforce. There is no meaningful “better exam” question here — there is only which platform pays your salary.

Side by side

	AB-620	Agentforce Specialist
Platform	Microsoft Copilot Studio	Salesforce
Heaviest section	Integrate and extend agents (40–45%)	AI Agents (35%)
Data content	Connectors and environment configuration	Data 360: identity resolution, harmonisation (20%)
Prerequisite knowledge	Power Platform familiarity	Salesforce platform and security model
Coding	Minimal	Minimal

More than the vendor difference suggests. Both exams test the same underlying design judgement:

CompTIA certification path

Sun, 20 Sep 2026 00:00:00 +0000

Take Network+ (N10-009) before any security certification — networking is the foundation security sits on, and skipping it leaves gaps that surface in every later exam. Security+ (SY0-801) is the natural next step; CompTIA expects it to launch on or around November 17, 2026, and until then SY0-701 is the version you can book. SecAI+ (CY0-001) is a specialist exam about securing AI systems and assumes you already work in security; it is not a next step from Network+.

Free AAIA sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

These follow ISACA’s style: scenario-led, from the auditor’s perspective, with a qualifier that decides the answer. Three of the five come from AI Operations, matching the real weighting.

Question 1. An auditor is reviewing a credit-scoring model. Management asserts the model was validated before release. Which is the BEST audit evidence of that assertion?

A. A written statement from the lead developer confirming validation
B. The model’s source code and training scripts
C. Dated validation results against defined acceptance thresholds, with evidence of approval
D. The vendor’s product documentation describing the model’s accuracy

Show answer

Answer: C

Free AAIR sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

These follow ISACA’s style: scenario-led, from the risk professional’s perspective, with a qualifier that decides the answer. The mix reflects the weighting, so most come from governance and programme management.

Question 1. An enterprise is preparing to adopt AI across several business units. No AI systems have yet been catalogued. Which should the risk practitioner recommend FIRST?

A. Select an AI risk framework and adopt it enterprise-wide
B. Establish an inventory of AI systems and their intended use cases
C. Write an AI acceptable use policy and circulate it
D. Deploy monitoring tooling across all AI workloads

Show answer

Answer: B

Free AAISM sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

These follow ISACA’s house style: scenario-led, from the accountable manager’s perspective, with a qualifier that decides the answer. Read the capitalised word carefully.

Question 1. An enterprise has discovered that several business units are using public generative AI tools with company data, without approval. Which should the security manager do FIRST?

A. Issue a policy prohibiting all public generative AI tools
B. Deliver awareness training to all staff
C. Identify and inventory the AI tools in use and the data involved
D. Block access to all public AI services at the network perimeter

Show answer

Answer: C

Free AB-620 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the real weighting, so two come from the integration domain.

Question 1. An agent must retrieve records from a legacy internal application that has no API, no connector and no supported integration surface. Users interact with it only through its web interface. Which Copilot Studio capability fits?

A. A custom connector
B. A REST API tool
C. Computer use
D. Azure AI Search as a knowledge source

Show answer

Answer: C

Free Agentforce Specialist sample questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so AI Agents appears most.

Question 1. An agent must always apply the same discount calculation and never vary its wording when quoting the figure to a customer. How should this be implemented?

A. Instruct the agent in its topic description to always phrase it the same way
B. Use deterministic mechanisms such as variables and template expressions
C. Select a model with a lower temperature setting
D. Remove the discount capability from the agent

Show answer

Answer: B

Free AI-103 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these questions before you look at the answers. Each one maps to an official AI-103 skill area.

Question 1. Your agent in Microsoft Foundry must call an Azure AI Search index. Company policy forbids storing keys in application code or configuration. What should you use to authenticate?

A. An admin key stored in an environment variable
B. A managed identity with a role assignment on the search service
C. A query key hard-coded in the agent definition
D. A shared access signature (SAS) token

Show answer

Answer: B

Free AI-901 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before you look at the answers. Each maps to an official AI-901 skill area, and the mix reflects the real weighting: two on concepts, three on implementation.

Question 1. A bank uses an AI model to screen loan applications. Testing shows it approves applicants from one postal district at a noticeably lower rate than equally qualified applicants elsewhere. Which responsible AI principle does this primarily concern?

A. Fairness
B. Reliability and safety
C. Transparency
D. Accountability

Show answer

Answer: A

Free AIF-C01 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix reflects the real weighting, so most come from the generative AI domains.

Question 1. A company wants to predict next quarter's sales figures from five years of historical sales data. Which type of machine learning problem is this?

A. Classification
B. Regression
C. Clustering
D. Reinforcement learning

Show answer

Answer: B

Predicting a continuous numeric value such as a sales figure is regression. Classification predicts a category, clustering groups similar records without labels, and reinforcement learning trains an agent through rewards. The historical data with known outcomes also makes this supervised learning.

Free CY0-001 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix reflects the real weighting, so two of the five come from securing AI systems.

Question 1. An attacker submits crafted records into the dataset used to retrain your fraud-detection model, so that transactions matching a pattern they control are later classified as legitimate. What is this attack called?

A. Model extraction
B. Data poisoning
C. Prompt injection
D. Denial of service

Show answer

Answer: B

Free Databricks Data Analyst Associate questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so querying, dashboards and Genie appear most.

Question 1. An analyst joins an orders table to a customers table and the row count increases from 50,000 to 73,000. Each order should appear once. What is the most likely cause?

A. A missing WHERE clause on the orders table
B. The SQL warehouse is undersized for the join
C. The join key is not unique in the customers table, so orders match multiple rows
D. An inner join was used instead of a left join

Show answer

Answer: C

Free Databricks Data Engineer Associate questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so ingestion and transformation appear most.

Question 1. A nightly job reprocesses an entire cloud storage location even though only a small number of new files arrive each day. Cost and runtime are both rising. What should the engineer implement?

A. A larger cluster to reduce runtime
B. Incremental ingestion that processes only newly arrived files
C. Convert the source files to a different format
D. Run the job weekly instead of nightly

Show answer

Answer: B

Free Databricks Data Engineer Professional questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so code and optimisation appear most. Question 4 now falls under Data Ingestion & Acquisition, which absorbed sharing and federation in October 2026.

Question 1. A PySpark job joins a 2 TB fact table with a 40 MB dimension table. The job spends most of its time shuffling. What is the most effective change?

A. Increase the cluster size
B. Broadcast the small dimension table to avoid shuffling the fact table
C. Repartition the fact table before the join
D. Cache the fact table in memory

Show answer

Answer: B

Free Databricks GenAI Engineer sample questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the real weighting, so most come from development and deployment.

Question 1. A RAG application returns fluent answers that cite the wrong section of a long technical manual. Retrieval returns five chunks per query. What is the most likely cause?

A. The model is too small for technical content
B. Chunking strategy — chunks are too large or lack overlap, so retrieval matches poorly
C. The temperature setting is too high
D. The maximum output length is too short

Show answer

Answer: B

Free Databricks ML Associate sample questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so most come from the platform and model development sections.

Question 1. A data scientist has run twenty training experiments with different hyperparameters and needs to compare their metrics and recover the exact parameters that produced the best result. Which Databricks capability provides this?

A. MLflow tracking
B. Unity Catalog
C. Delta Live Tables
D. A cluster policy

Show answer

Answer: A

Free Databricks ML Professional sample questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. Four come from the two 44% sections, matching the real weighting.

Question 1. A model validated at 0.94 AUC performs at 0.71 in production from day one. Investigation finds a feature derived from a field populated only after the outcome is known. What has occurred?

A. Concept drift
B. Overfitting to the training set
C. Data leakage — the feature is unavailable at prediction time
D. Insufficient model capacity

Show answer

Answer: C

Free MLA-C02 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. One from each domain, plus an extra from data preparation to match the weighting.

Question 1. A model trained on transaction data performs far better in testing than in production. Investigation shows a column derived from the final transaction status was included as a feature. What has happened?

A. Overfitting caused by too many training epochs
B. Data leakage — a feature encodes information unavailable at prediction time
C. Model drift since deployment
D. Insufficient endpoint compute capacity

Show answer

Answer: B

Free N10-009 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so troubleshooting and concepts appear most.

Question 1. A user can reach websites by IP address but not by name. Other users on the same subnet are unaffected. Which is the most likely cause?

A. The default gateway has failed
B. The patch cable to the workstation is faulty
C. The workstation’s DNS configuration is incorrect
D. A duplex mismatch on the switch port

Show answer

Answer: C

Free SC-500 sample questions with answers

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. One comes from each domain, plus an extra from the new AI security material.

Question 1. Administrators need standing access removed. They should request elevation to a privileged role, be approved, and hold it only for a limited period, with an audit trail. What should you implement?

A. A permanent Owner role assignment for each administrator
B. Privileged Identity Management with approval and time-bound activation
C. A conditional access policy requiring MFA
D. A resource lock on the subscription

Show answer

Answer: B

Free Spark Developer Associate sample questions

Sun, 20 Sep 2026 00:00:00 +0000

Try these before reading the answers. The mix follows the weighting, so the DataFrame API, architecture and Spark SQL appear most.

Question 1. Which of these transformations causes a shuffle?

A. filter
B. select
C. groupBy
D. withColumn

Show answer

Answer: C

groupBy requires records with the same key to be co-located, so data moves across the cluster, making it a wide transformation. filter, select and withColumn each operate within a partition without moving data and are narrow transformations.

200-901 software development explained

Fri, 09 Oct 2026 00:00:00 +0000

Software Development and Design is worth 15% of 200-901. It covers data formats and how they parse into Python, test-driven development, development methods, code organisation, two design patterns and the Git operations you use daily. The questions are conceptual and short, and they favour candidates who have actually committed code and resolved a merge conflict.

Data formats and Python structures

Format	Style	Python result
JSON	Braces, brackets, quoted keys	`dict`, `list`, `str`, `int`, `bool`, `None`
YAML	Indentation, dashes for lists	Same structures; indentation errors are common
XML	Tags and attributes	A tree of elements; parsed with `xml.etree.ElementTree`

JSON true, false and null become True, False and None. json.loads takes a string; json.load takes a file. YAML is a superset of JSON in practice and is favoured for configuration such as Ansible playbooks; load it with yaml.safe_load. XML is verbose, supports attributes and namespaces, and is still used by NETCONF and some Cisco APIs.

AB-100 Plan AI solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

The Plan AI-powered business solutions domain is 25–30% of AB-100 and tests whether you can decide what to build before anyone builds it. It has three parts: analysing requirements and data, designing the overall AI strategy, and evaluating costs and benefits. The questions describe a business situation and ask which approach, platform or investment case is the most sensible.

Analyse requirements

The official outline asks you to assess where agents help with task automation, data analytics and decision-making, to review data used for grounding, and to organise business data so other AI systems can use it.

AB-730 generative AI fundamentals explained

Fri, 09 Oct 2026 00:00:00 +0000

Understand generative AI fundamentals is the largest AB-730 domain, worth 25–30% of the exam. It tests whether you know how Copilot arrives at an answer, what influences that answer, which Copilot experience suits a task, and what can go wrong. It is a business-user view, not a machine learning course: expect scenarios in which you pick the safest or most sensible action.

The outline has two groups: generative AI capabilities across Microsoft 365 experiences, and responsible AI and data protection practices.

AB-731 generative AI business value explained

Fri, 09 Oct 2026 00:00:00 +0000

Identify the business value of generative AI solutions is worth 35–40% of AB-731, and it is the exam’s vocabulary domain. It asks you to explain how generative AI differs from other AI, what drives its cost, what makes it reliable or unreliable, and when it actually pays off. You are not asked to build anything, but you must reason like someone who signs off budgets.

Generative AI versus other AI

Type	What it does	Example
Machine learning (predictive)	Learns patterns from data to predict or classify	Forecast demand, flag likely churn
Generative AI	Produces new content: text, images, code, summaries	Draft a proposal, summarise a meeting
Rules and automation	Follows fixed logic	Route an invoice above a set amount

Choose generative AI when the task involves language or content, variation is acceptable and humans can review the output. Choose predictive ML when you need a number or a class from structured history. Choose plain automation when the logic is fixed and must be exact. The outline asks you to select a solution for a business need, so practise naming the right type before the right product.

AZ-305 identity, governance and monitoring

Fri, 09 Oct 2026 00:00:00 +0000

Design identity, governance, and monitoring solutions is worth 25–30% of AZ-305. It covers three areas: recommending how logs are collected and routed, recommending how people and workloads authenticate and are authorized, and recommending how an estate of subscriptions is structured and kept compliant. The questions are about choosing the right model for an organisation, not about clicking through a setting.

Design solutions for logging and monitoring

Logs and where they go

Azure Monitor collects metrics and logs; Log Analytics workspaces store the logs and answer KQL queries. A diagnostic setting on a resource sends its logs to one or more destinations:

AZ-400 processes and communications explained

Fri, 09 Oct 2026 00:00:00 +0000

Design and implement processes and communications is 10-15% of AZ-400. It covers how work is tracked from idea to release, which metrics tell you whether delivery is healthy, and how teams document and talk about their work. The questions are less technical than the pipeline domain and are often won by knowing which tool links to which: Azure Boards, GitHub Issues and Projects, wikis, webhooks and Microsoft Teams.

Traceability and flow of work

The outline names three sub-topics here.

CCNA 200-301 network infrastructure explained

Fri, 09 Oct 2026 00:00:00 +0000

Network Infrastructure and Connectivity is worth 25% of the v2.0 CCNA, tied for the largest domain. It covers physical interface and cable faults, hypervisors and containers, IPv4 and IPv6 troubleshooting including subnetting, wireless principles, client connectivity and DHCPv4. The verb throughout is troubleshoot: you are shown a symptom and asked for the cause.

Interfaces and cabling

Collisions, errors, a duplex or speed mismatch, distance limits, signal levels, pinouts and cable types are all named in the objectives.

DP-600 maintain a solution explained

Fri, 09 Oct 2026 00:00:00 +0000

Maintain a data analytics solution is worth 25–30% of DP-600. It has two halves: security and governance (who can see what, and how items are labelled and endorsed) and the analytics development lifecycle (how changes move from a developer’s machine to production). The questions are short and factual, and they reward knowing which control works at which level.

Implement security and governance

Layers of access

Fabric security is layered. A request has to pass each layer that applies, and you pick the layer that matches the requirement.

DP-700 implement and manage explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement and manage an analytics solution is worth 30–35% of DP-700. It covers four areas: workspace settings, lifecycle management, security and governance, and orchestration. The questions are rarely about code. They ask which setting, role or item type solves a described requirement, so the skill is matching the requirement to the right Fabric feature.

Configure Fabric workspace settings

The outline names Spark, domain, OneLake and Apache Airflow workspace settings.

Spark settings control the pool that notebooks and Spark job definitions use, the runtime version and session behaviour. An Environment item bundles libraries, Spark properties and compute settings so that many notebooks share the same configuration.
Domains group workspaces by business area so governance can be delegated.
OneLake settings include whether external apps can read OneLake data and how OneLake security is used.
Apache Airflow settings apply to Apache Airflow jobs, which run Python DAGs for orchestration.

Lifecycle management

Capability	What it does
Git integration	Connects a workspace to a repository (Azure DevOps or GitHub) so items are version controlled, with branches and commits
Database projects	Keep a warehouse schema as code so it can be built and deployed from source
Deployment pipelines	Move content through development, test and production stages
Deployment rules	Override settings per stage, such as data source, parameters or a notebook’s default lakehouse

A typical question describes three stages and asks how to stop test items pointing at development data. That is a deployment rule, not a manual edit.

DP-750 environment and compute explained

Fri, 09 Oct 2026 00:00:00 +0000

Set up and configure an Azure Databricks environment is worth 15–20% of DP-750. It has two halves: choosing and configuring compute, and creating the Unity Catalog objects your data lives in. The questions are choice-based: given a workload and a constraint, which compute type, setting or object type fits. If you can say why a nightly job should not run on an all-purpose cluster, you already understand half of this domain.

DP-800 design and develop explained

Fri, 09 Oct 2026 00:00:00 +0000

Design and develop database solutions is the joint-largest DP-800 domain at 35–40%. It covers four areas: designing database objects, building programmability objects, writing advanced T-SQL, and using AI-assisted tools such as GitHub Copilot and MCP servers. Most questions show a schema or a query and ask what it does, what is wrong with it, or which feature fits a requirement.

Design and implement database objects

Tables, indexes and columnstore

Choose the smallest data type that fits the data, because width affects index size, memory and I/O. A rowstore clustered index suits transactional lookups; a columnstore index suits large scans and aggregates and compresses well. A nonclustered columnstore index on a transactional table gives you analytics on live data.

PL-300 prepare the data explained

Fri, 09 Oct 2026 00:00:00 +0000

Prepare the data is worth 25–30% of PL-300, the joint-largest domain. It covers three things: connecting to sources and choosing how the data is stored, profiling and cleaning what arrives, and transforming it into tables the model can use. Almost all of it happens in the Power Query editor, and the questions test whether you know which transformation fits a situation, and what it does to rows, columns and refresh.

SC-200 security operations environment

Fri, 09 Oct 2026 00:00:00 +0000

Manage a security operations environment is worth 40–45% of SC-200, the largest domain. It covers configuration rather than investigation: automating Defender XDR and Sentinel, setting up the Sentinel platform, getting data into it, and building the detections that raise incidents. The questions ask which setting, rule type or connector fits a stated requirement, so you need to know what each feature does and what it cannot do.

Configure automation for Defender XDR and Sentinel

Defender XDR and Defender for Endpoint

The objectives cover email and alert notifications (including tuning, suppression and correlation), advanced features and rule settings in Defender for Endpoint, custom data collection, security policies including attack surface reduction (ASR) rules, automated investigation and response, automatic attack disruption, and device groups with permissions and automation levels.

SC-300 user identities explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement and manage user identities is worth 20–25% of SC-300. It covers four things: configuring the Microsoft Entra tenant (roles, administrative units, domains, settings), creating and managing identities, handling external users and other tenants, and connecting an on-premises directory. Most questions ask which tool, role or sync option fits a stated requirement.

Configure and manage the tenant

Roles and administrative units

Entra has built-in roles and, with the right licence, custom roles made from selected permissions. Roles are assigned at tenant scope, or scoped to an administrative unit (AU), a container of users, groups or devices. A scoped role applies only to the members of that AU. The classic scenario is a regional helpdesk that resets passwords only for its own staff. AUs can have static membership or dynamic membership rules. You are also expected to evaluate effective permissions: a user can hold several role assignments, and the result is the union of all of them.

SC-401 information protection explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement information protection is worth 30–35% of SC-401. It has three parts: classifying data so Microsoft Purview can recognise it, creating and publishing sensitivity labels that protect it, and extending protection to Windows, file shares and email. The questions ask which classification method or label setting fits a requirement, and what has to be in place for it to work.

Data classification

Sensitive info types and how to choose

Method	Use it when
Built-in sensitive info type	The data is a standard format such as a credit card or national ID number
Custom sensitive info type	The data follows your own pattern: regular expression, keywords, optional checksum and supporting evidence
Document fingerprinting	You want to detect documents built from a standard form or template
Exact data match (EDM)	You must match actual values from a structured table, such as customer records, with very few false positives
Trainable classifier	The content is recognisable by its meaning, not a pattern, such as contracts or source code

EDM works by hashing the sensitive table with a salt and uploading only the hashes, so the raw data never reaches the cloud. Trainable classifiers learn from a seed set of sample documents you provide; Microsoft also supplies pre-trained ones. Optical character recognition lets sensitive info types detect text inside images.

SC-900 security, compliance and identity concepts explained

Fri, 09 Oct 2026 00:00:00 +0000

The concepts area is the smallest SC-900 skill area at 10–15%, and it supplies the vocabulary for the rest of the exam. It covers five security and compliance ideas (shared responsibility, defense in depth, Zero Trust, encryption and hashing, and governance, risk and compliance) and the core identity ideas: identity as the security perimeter, authentication, authorization, identity providers, directory services and federation. Questions here are short definitions or scenarios where you pick the right term.

SY0-801 general security concepts explained

Fri, 09 Oct 2026 00:00:00 +0000

General Security Concepts is worth 16% of SY0-801. It has three objectives: security concepts and controls (1.1), the effect of change management on security (1.2), and appropriate cryptographic solutions (1.3). It is the shortest domain by objective count but the vocabulary appears in every other domain, so errors here cost marks everywhere.

1.1 Security concepts and controls

The core concepts are defence in depth, the CIA triad (confidentiality, integrity, availability), AAA (authentication, authorization, accounting), non-repudiation, Zero Trust principles and least privilege.

Sysadmin to cloud: Azure and AWS certifications

Fri, 09 Oct 2026 00:00:00 +0000

For a system administrator, the cloud move on this site runs through Azure: AZ-104 (Azure Administrator Associate) is the main exam, AZ-802 is the alternative if the work stays mostly Windows Server, and AZ-305 follows for design. On the AWS side the site covers only the AI exams, AIF-C01 and MLA-C02; it has no guide to an AWS infrastructure or administrator exam, and this page does not pretend otherwise.

AZ-802 AD DS explained

Tue, 29 Sep 2026 00:00:00 +0000

Deploy and manage AD DS is the largest AZ-802 domain at 20–25%. It covers four areas: deploying domain controllers, running multi-site and multi-forest environments, managing users, groups and service accounts, and managing servers with Group Policy. The questions are practical. You need to know which FSMO role explains a symptom, which trust type fits two forests, and which service account type a given workload can use.

Domain controllers

On-premises and in Azure

Promoting a DC is the easy part. The exam cares about the decisions around it. A domain controller in an Azure VM should have a static private IP, store the AD database, logs and SYSVOL on a data disk with host caching set to None, and point its DNS settings at domain controllers rather than Azure’s resolver. Put Azure DCs in their own AD site, with the Azure address ranges as subnets, so clients there authenticate locally.

AB-900 Microsoft 365 core features explained

Mon, 28 Sep 2026 00:00:00 +0000

Identify the core features and objects of Microsoft 365 services is worth 30–35% of AB-900. It has three parts: the objects each admin center manages, the security principles behind Microsoft 365, and the identity features that enforce them in Microsoft Entra ID. The skill being tested is knowing where things live and which tool answers a given question — not configuring them in depth.

Which admin center owns what

Many questions describe a task and ask where to do it.

AI-300 MLOps infrastructure explained

Mon, 28 Sep 2026 00:00:00 +0000

Design and implement an MLOps infrastructure is worth 15–20% of AI-300. It covers the ground everything else stands on: the Azure Machine Learning workspace and its resources, the assets you reuse across jobs, and the infrastructure as code and GitHub Actions workflows that create them. Questions here are about setting things up correctly and securely, so that training and deployment can be automated later.

Workspace resources

A workspace is the top-level container for experiments, models, endpoints and assets. Creating one also links, or creates, a storage account, a Key Vault, Application Insights and optionally a container registry. Knowing which dependent resource does what answers a surprising number of questions: images for environments go to the registry, secrets to Key Vault, endpoint telemetry to Application Insights.

AZ-104 identities and governance explained

Mon, 28 Sep 2026 00:00:00 +0000

Manage Azure identities and governance is one of the two largest AZ-104 domains at 20–25%. It covers three things: users and groups in Microsoft Entra ID, access to Azure resources through role-based access control, and the governance tools that keep subscriptions tidy and affordable — Azure Policy, locks, tags, management groups, budgets and Advisor. Most questions are about scope: who gets what, and where it applies.

Microsoft Entra users and groups

Users and groups

You create users in the portal, with PowerShell or the CLI, or in bulk from a CSV file. Groups come in two types, security and Microsoft 365, and two membership styles:

AZ-900 cloud concepts explained

Mon, 28 Sep 2026 00:00:00 +0000

Describe cloud concepts is the AZ-900 domain worth 25–30% of the exam. It is the only part that is not specific to Azure: it tests whether you understand what cloud computing is, who is responsible for what, which cloud model and service type fit a situation, and why organisations move to the cloud at all. Most of it is vocabulary, but the questions test whether you can apply that vocabulary to a short scenario.

AI-200 containerized solutions explained

Sun, 27 Sep 2026 00:00:00 +0000

Develop containerized solutions on Azure is worth 20–25% of AI-200. It covers two things: hosting a container (building and storing images in Azure Container Registry, then running them on App Service) and orchestrating containers (Azure Container Apps and Azure Kubernetes Service). The questions reward knowing which platform fits a scenario, and reading the configuration that makes it work.

Container application hosting

Azure Container Registry

The registry is where every other objective in this domain starts. Know how to:

SEE Part 1 preliminary work explained

Mon, 21 Sep 2026 00:00:00 +0000

Preliminary Work and Taxpayer Data accounts for 14 of the 85 scored questions on SEE Part 1, about 16%. It covers everything a preparer settles before calculating any tax: who the taxpayer is, whether they must file, which filing status and dependents they may claim, their residency status, and which other returns and reporting obligations apply. These are the facts that every later answer depends on.

The questions tend to be fact-heavy rather than computational. A scenario describes a household, and you decide what follows from it.

SEE Part 2 business entities explained

Mon, 21 Sep 2026 00:00:00 +0000

Business Entities and Considerations is 30 of the 85 scored questions on SEE Part 2, about 35%. It tests how each type of entity is classified, formed and taxed, and how money and property move between the entity and its owners. The common thread is basis: an owner’s basis decides what is taxable when cash comes out and which losses can be deducted.

Entity types and classification

Entity	Default treatment	Return
Sole proprietorship	Not separate from the owner	Schedule C with Form 1040
Single-member LLC	Disregarded entity	Owner’s return (e.g. Schedule C)
Multi-member LLC	Partnership	Form 1065
General or limited partnership	Partnership	Form 1065
Corporation	C corporation	Form 1120
Corporation or LLC with an S election	S corporation	Form 1120-S

An eligible LLC can elect to be taxed as a corporation on Form 8832; an eligible corporation, or an LLC treated as one, elects S status on Form 2553. A qualified joint venture lets a married couple who jointly own and operate an unincorporated business, both materially participate and file jointly, skip Form 1065 and each report their share on their own Schedule C.

SEE Part 3 practices and procedures explained

Mon, 21 Sep 2026 00:00:00 +0000

Practices and Procedures is the largest SEE Part 3 domain, with 26 of the 85 scored questions (about 31%). It covers two bodies of rules: Circular 230, the Treasury regulations governing practice before the IRS (31 CFR Part 10), and the Internal Revenue Code penalties on return preparers. Read Circular 230 itself; the exam uses its wording.

What practice before the IRS means

Practice covers any presentation to the IRS about a taxpayer’s rights, privileges or liabilities: filing documents, corresponding with the IRS, written advice, and representing a client at conferences and hearings.

AAIA AI governance and risk explained

Sun, 20 Sep 2026 00:00:00 +0000

AI Governance and Risk is worth 33% of AAIA — around 30 of the 90 questions. ISACA breaks it into five areas, and for an auditor the useful framing is: what should exist, who owns it, and what would prove it exists.

AI models, considerations and requirements

You cannot audit what you cannot describe. This area asks for working literacy:

What a model is, and that its behaviour derives from training data rather than written logic
Training versus inference — two distinct stages with distinct risks and distinct evidence
Why outputs are probabilistic, so “correct” is a threshold rather than a binary
What requirements an AI solution should have defined before it was built: intended use, acceptable performance, limitations, and who it affects

The audit angle: requirements defined after deployment are not requirements. A scenario where nobody can state the intended use is a governance finding, not a technical one.

AAIR AI risk governance and frameworks

Sun, 20 Sep 2026 00:00:00 +0000

AI Risk Governance and Framework Integration is worth 37% of AAIR — around 33 of the 90 questions. ISACA splits it into six areas, and the through-line is integration: AI risk governance should extend what the enterprise already has, not sit beside it as a parallel structure.

AI models, frameworks, strategies and use cases

Working literacy, not technical depth:

Training versus inference, and why behaviour derives from data rather than written logic
Why outputs are probabilistic, so “acceptable” is a threshold not a binary
That use cases carry very different risk profiles — an internal drafting assistant and a credit decisioning model are not comparable

The practical skill is classifying a use case by risk, because almost everything downstream is proportionate to that classification.

AAISM AI governance and program management

Sun, 20 Sep 2026 00:00:00 +0000

AI Governance and Program Management is worth 31% of AAISM — around 28 of the 90 questions. ISACA describes it as advising stakeholders on implementing AI security solutions through appropriate and effective policy, data governance, programme management and incident response.

If you hold CISM, the discipline is yours already. What is new is everything AI-specific inside it, and the temptation to skim this domain because the headings look familiar is the main risk.

AAISM vs SecAI+ (CY0-001)

Sun, 20 Sep 2026 00:00:00 +0000

Both cover securing AI systems, and they are aimed at different seats in the room. AAISM is for security managers and tests governance, programme management and control selection. CY0-001 is for practitioners and tests attacks, defences and hands-on reasoning. Pick the one matching what you are asked to produce: a decision, or a configuration.

Side by side

	AAISM	CY0-001
Body	ISACA	CompTIA
Audience	Security managers	Security practitioners
Heaviest domain	AI Technologies and Controls (38%)	Securing AI systems (40%)
Governance weight	31% governance + 31% risk	19% governance, risk and compliance
Assumes	Security management experience	Working security experience

The distinction that matters

The domain names look similar. The questions do not.

AB-620 plan and configure agent solutions

Sun, 20 Sep 2026 00:00:00 +0000

Plan and configure agent solutions is worth 30–35% of AB-620. It splits into three unequal parts: planning decisions you make before building, agent flows, and topic configuration. The planning part rewards judgement; the other two reward having actually built something in Copilot Studio.

Planning an agent solution

Seven planning objectives, each of which is really a design question the exam poses as a scenario:

Integration with enterprise systems. Which mechanism reaches which system — connector, custom connector, REST API, MCP, or computer use when nothing else exists.
Identity strategy. Who the agent acts as. The recurring distinction is whether the agent uses the user’s identity, so results respect that person’s permissions, or a single service identity that sees everything. Any scenario mentioning “users should only see what they are allowed to see” wants the former.
Channels and deployment. Where the agent is surfaced — Teams, a website, Microsoft 365 Copilot — and what that implies for authentication and audience.
Responsible AI strategy. Content controls, oversight, and what the agent must refuse.
Security and governance. Data boundaries, environment strategy, who can publish.
Reusable agent components. Build once, use in several agents.
Internal versus external audiences. An externally facing agent needs tighter guardrails, authentication decisions and a narrower scope than an internal one.

Agent flows

An agent flow is the deterministic part of an agent — the bit that runs steps reliably rather than deciding conversationally. The objectives call out:

Agentforce Specialist prompt engineering

Sun, 20 Sep 2026 00:00:00 +0000

Prompt Engineering is worth 20% of the Salesforce Agentforce Specialist exam — around 12 of the 60 scored questions. It covers Prompt Builder, the templates you create in it, and the Trust Layer that sits between your data and the model.

When Prompt Builder is appropriate

The first judgement the exam tests. Prompt Builder suits generating content from Salesforce data — a summary, a draft email, a field value. It is not the answer when the requirement is to do something, which is an agent action, or to run deterministic logic, which is a flow.

AI-103 plan and manage an Azure AI solution explained

Sun, 20 Sep 2026 00:00:00 +0000

Plan and manage an Azure AI solution is worth 25–30% of your AI-103 score, and it is the domain candidates underestimate most. It contains very little coding. Instead it asks whether you can pick the right model and service for a job, deploy it sensibly, keep it secure and inside its budget, and show that it behaves responsibly. If you come from a developer background, this is usually where you lose points.

AI-901 identify AI concepts and capabilities explained

Sun, 20 Sep 2026 00:00:00 +0000

Identify AI concepts and capabilities is worth 40–45% of AI-901 and it is the half you can genuinely learn from reading. It has three parts: the principles of responsible AI, how models and their configuration work, and recognising which AI workload a scenario calls for. None of it requires you to build anything — but it does require precision, because the questions are built on distinctions that sound similar until you look closely.

AIF-C01 fundamentals of AI and ML explained

Sun, 20 Sep 2026 00:00:00 +0000

Fundamentals of AI and ML is worth 20% of AIF-C01’s scored content — around 10 of the 50 scored questions. It is the vocabulary layer, and everything in the other four domains is expressed in it. Get this wrong and you will misread questions elsewhere.

The nesting

Learn this as concentric circles, because questions test the distinction:

Artificial intelligence — the broad field of systems performing tasks that need human-like intelligence.
Machine learning — a subset where systems learn patterns from data rather than following written rules.
Deep learning — a subset of ML using multi-layered neural networks, behind most modern vision and language capability.
Generative AI — applications producing new content, built on deep learning.

The three learning types

Type	Data	Typical use
Supervised	Labelled — inputs with known answers	Classification, regression
Unsupervised	Unlabelled	Clustering, anomaly detection, grouping
Reinforcement	Rewards from an environment	Sequential decisions, robotics, game playing

The exam’s usual framing: a scenario describes the data available, and you identify the learning type. If the historical data includes known outcomes, it is supervised. If you are looking for structure nobody has labelled, it is unsupervised.

CY0-001 basic AI concepts related to cybersecurity explained

Sun, 20 Sep 2026 00:00:00 +0000

Basic AI concepts related to cybersecurity is worth 17% of CY0-001 — around ten questions. It is the smallest domain and the one most people skim, which is a mistake, because everything in the other three domains is expressed in this vocabulary. Get it wrong here and you will misread questions elsewhere.

The lifecycle, and where attacks land

Learn the stages, then learn which threat belongs to which stage. That single mapping answers a surprising share of the exam.

Databricks Data Analyst: platform and data

Sun, 20 Sep 2026 00:00:00 +0000

Understanding of the Databricks Data + AI Platform (11%), Managing Data (8%) and Importing Data (5%) are 24% of the Data Analyst Associate exam between them — around 11 of the 45 scored questions.

These are the foundations. None is large individually, but together they are nearly a quarter of the paper, and they underpin every other section.

Understanding the platform (11%)

The lakehouse idea — one storage layer serving analytics and engineering, rather than a warehouse kept in sync with a lake.
Delta tables as the default format, and what that gives you: ACID transactions, schema enforcement, time travel.
How the workspace is organised: catalogs, schemas, tables, dashboards, queries.
Where Databricks SQL sits — the analyst-facing surface.

SQL warehouses deserve particular attention:

Databricks Data Engineer: platform and ingestion

Sun, 20 Sep 2026 00:00:00 +0000

Databricks Intelligence Platform (6%) and Data Ingestion and Loading (21%) are 27% of the Data Engineer Associate exam between them — around 12 of the 45 scored questions. The platform section is tiny; ingestion is one of the two largest on the exam.

The platform section (6%)

About three questions. Cover it in an evening:

The lakehouse idea: one storage layer serving both analytics and data engineering, rather than a warehouse and a lake kept in sync.
Delta Lake as the default table format — ACID transactions, schema enforcement, and time travel over files in cloud storage.
Workspace organisation: notebooks, repos, clusters, catalogs.
Compute basics: which compute type fits which workload, when it starts and stops, and why leaving it running costs money (details below).

That is genuinely enough for a 6% section. Do not spend a week here.

Databricks DE Professional: developing code

Sun, 20 Sep 2026 00:00:00 +0000

Developing Code for Data Processing using Python and SQL is Section 1 of the Databricks Data Engineer Professional exam and worth 23% in the outline that applies from 9 October 2026: around 14 of the 60 scored questions and the largest section. It has no counterpart on the Associate, and it is the clearest marker of the step up.

The exam is multiple choice, so you read code rather than write it. But the fluency required is real: you will be shown a snippet and asked what it produces, what is wrong with it, or which of four versions is correct.

Databricks GenAI design and data preparation

Sun, 20 Sep 2026 00:00:00 +0000

Design Applications and Data Preparation are 14% each — 28% of the Databricks Generative AI Engineer Associate exam between them, roughly 12 of the 45 scored questions. They cover the decisions made before any application code exists, and the data work that determines whether it will ever work well.

Designing a generative AI application

The first decision is what approach the requirement calls for:

Approach	Use when
Prompting alone	The model already knows enough; you need format or tone control
RAG	Answers must come from your content, especially content that changes
Fine-tuning	You need consistent behaviour, style or domain vocabulary
Chains and agents	The task needs multiple steps or tool use

The rule that decides most exam questions: if the information changes, it is RAG. If the behaviour needs to change, it is fine-tuning. A scenario mentioning documents updated weekly, current data, or citing sources is a RAG question every time.

Databricks ML Associate: the platform section

Sun, 20 Sep 2026 00:00:00 +0000

Databricks Machine Learning is worth 38% of the Machine Learning Associate exam — around 18 of the 48 scored questions and the largest section by a clear margin. It is platform knowledge, not modelling theory, and it is where candidates with strong ML backgrounds most often lose marks.

If you learn one section properly, learn this one.

The ML runtime and workspace

The Databricks ML runtime, which ships the common ML libraries preinstalled, and why you would choose it over a standard runtime.
Cluster configuration for ML work, including when a GPU is justified.
Notebooks and repos as the working surface.

MLflow: tracking

The component that appears most often on this exam.

Databricks ML Professional: model development

Sun, 20 Sep 2026 00:00:00 +0000

Model Development is worth 44% of the Databricks Machine Learning Professional exam — around 26 of the 59 scored questions, tied with ML Ops as the largest section.

It is not the same material as the ML Associate’s development section. That one asks whether you can train a model. This asks whether you can do it rigorously: validated properly, evaluated honestly, and reproducible months later.

Validation strategy

The foundation, and where professional-level questions concentrate.

ISACA AI certification path

Sun, 20 Sep 2026 00:00:00 +0000

Pick the one that matches your existing role: AAIA if you audit, AAIR if you manage risk, AAISM if you manage security. None of the three is an entry-level credential, despite the AI framing — they assume you already work in governance, risk or assurance and are extending into AI.

The three certificates

Exam	For	Core weighting
AAIA	Auditors	AI Operations 46%, AI Governance and Risk 33%, AI Auditing Tools and Techniques 21%
AAIR	Risk professionals	AI Risk Program Management 42%, AI Risk Governance 37%, AI Life Cycle Risk 21%
AAISM	Security managers	AI Technologies and Controls 38%, AI Governance 31%, AI Risk Management 31%

They are siblings, not a sequence

This is the most common misunderstanding. AAIA, AAIR and AAISM are not three levels of the same thing. They are three professional lenses on the same subject, and the right one is determined by your job title rather than your experience level.

MLA-C02 data preparation for ML and AI

Sun, 20 Sep 2026 00:00:00 +0000

Data Preparation for ML and AI is worth 28% of MLA-C02 — the largest domain. AWS frames the task as ingesting, transforming, validating and preparing data for AI and ML modelling.

Engineers arriving from a modelling background routinely underrate this. In practice most ML failures are data failures, and AWS has weighted the exam to reflect that.

Ingestion and storage

Getting data in, and choosing where it lives:

The common AWS sources and how data arrives — batch versus streaming.
Storage choices, and the cost and access trade-offs behind them.
Data formats. Columnar formats compress well and read efficiently for analytics; row formats suit record-at-a-time access. Choosing the right one is a cost and performance decision the exam tests.
Partitioning, so queries and training jobs read only what they need.

Transformation and cleaning

Handling missing values, duplicates and outliers, and knowing that each choice is a modelling decision rather than housekeeping.
Type conversion, normalisation and scaling.
Joining and aggregating across sources.
Doing this at scale, which is what distinguishes this exam from a data science course.

Feature engineering

Creating features that carry signal, and encoding categorical variables.
Storing and reusing features so training and inference compute them identically.

That last point matters more than it sounds: training/serving skew, where features are computed one way in training and another at inference, is a classic production failure and a natural exam scenario.

N10-009 networking concepts explained

Sun, 20 Sep 2026 00:00:00 +0000

Networking concepts is worth 23% of CompTIA Network+ N10-009 — around 21 of the 90 questions and the second largest domain. It is also the domain that underpins the other four: addressing, the OSI model and protocols appear throughout implementation, operations, security and troubleshooting.

Weakness here costs far more than 23%.

The OSI model

The framework the whole exam thinks in.

Layer	Name	Lives here
7	Application	HTTP, DNS, SMTP
6	Presentation	Encoding, encryption
5	Session	Session establishment
4	Transport	TCP, UDP, ports
3	Network	IP, routers, routing
2	Data Link	MAC addresses, switches, VLANs
1	Physical	Cables, connectors, signals

Its practical value is diagnostic. Given a symptom, ask which layer it points at: no link light is layer 1, a MAC or VLAN problem is layer 2, wrong subnet or no route is layer 3, a blocked port is layer 4, name resolution is layer 7.

SC-500 identity, access and governance

Sun, 20 Sep 2026 00:00:00 +0000

Manage identity, access, and governance is worth 20–25% of SC-500. It has three parts: securing access with Microsoft Entra ID, protecting secrets with Azure Key Vault, and enforcing governance across the estate. Experienced engineers tend to be strong on the first two and casual about the third — which is where the marks leak.

Securing access with Entra ID

Six objectives, each with a clear scenario signature:

Objective	The scenario that wants it
Privileged Identity Management	Standing admin access must become time-bound and approved
Conditional access	Access depends on conditions — device, location, risk
MFA and passwordless	Authentication strength must increase
Identity for applications	An app needs its own identity (enterprise apps, app registrations)
OAuth permission grants and consent	Users are consenting to third-party apps unchecked
Managed identities	An Azure resource must authenticate without a stored secret

The distinction worth drilling: PIM controls how long you hold a role, conditional access controls the conditions under which you sign in. Questions frequently offer both, and only one matches the requirement.

Spark Developer: architecture and components

Sun, 20 Sep 2026 00:00:00 +0000

Apache Spark Architecture and Components is worth 20% of the Spark Developer Associate exam — around 9 of the 45 scored questions. That is unusually high for an associate certification, and deliberately so: without it, the tuning and troubleshooting content is unintelligible.

Study this section first. It explains everything else.

The execution model

Driver — runs your program, builds the execution plan, and coordinates work.
Executors — run tasks and hold data in memory across the cluster.
Cluster manager — allocates resources.

The practical consequence, tested repeatedly: anything that pulls data back to the driver — collect() on a large DataFrame — defeats distribution and can exhaust the driver’s memory.

200-901 APIs explained

Fri, 09 Oct 2026 00:00:00 +0000

Understanding and Using APIs is the heaviest single domain of 200-901 at 20%, and the skills spill into the platform and automation domains as well. It tests whether you can build a REST request from documentation, read the response, and work out what went wrong from a status code. Expect short code fragments and error scenarios.

The anatomy of a REST request

A request has a method, a URL (with path and query parameters), headers and an optional body. A response has a status code, headers and a body, usually JSON.

AB-100 Design AI solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

The Design AI-powered business solutions domain is 25–30% of AB-100 and asks you to choose the right kind of agent, the right Copilot Studio construct, and the right extension or prebuilt feature for a stated requirement. It has three parts: designing AI and agents, designing extensibility, and orchestrating the configuration of prebuilt agents and apps.

Design AI and agents

Choosing an agent type

The outline lists task agents, autonomous agents and prompt-and-response agents. Treat them as a ladder of independence:

AB-730 prompts and chats explained

Fri, 09 Oct 2026 00:00:00 +0000

Manage prompts and chats by using Microsoft 365 Copilot is worth 20–25% of AB-730 and is the most hands-on domain: how to write a good prompt, how to save, schedule and share it, how Copilot uses memory and instructions, and how to find, rename, delete and organise your chats and notebooks. If you use Copilot daily, much of this is muscle memory; the exam turns it into recall of exact options.

AB-731 Microsoft AI apps and services explained

Fri, 09 Oct 2026 00:00:00 +0000

Identify benefits, capabilities, and opportunities for Microsoft’s AI apps and services is worth 35–40% of AB-731. It is a product-matching domain: given a business need, you pick the Microsoft Copilot or Foundry capability that fits, and you justify whether to buy, extend or build. It has two parts, Microsoft 365 Copilot and Microsoft Copilot, and Foundry Tools.

Microsoft 365 Copilot and Copilot

Need	Best fit
General AI chat for work accounts, web research, with enterprise protection	Microsoft 365 Copilot Chat on web and mobile
Draft, summarise and analyse inside Word, Excel, PowerPoint, Outlook, Teams	Copilot in the Microsoft 365 apps
Answers grounded in my emails, meetings and files	Microsoft 365 Copilot, via Microsoft Graph
A multi-source research report across work and web data	Researcher
Step-by-step analysis of spreadsheets and data files	Analyst
A custom agent for a team or process, low code	Microsoft Copilot Studio

Microsoft Graph is the layer that connects Copilot to a person’s work data (mail, calendar, meetings, files, chats) while respecting that person’s permissions. That is why the same question can give two colleagues different answers, and why oversharing in SharePoint becomes a Copilot problem.

AZ-305 data storage solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

Design data storage solutions is worth 20–25% of AZ-305. It covers relational data, semi-structured and unstructured data, and data integration and analysis. The exam rarely asks how a database works internally. It describes data, access patterns, scale and budget, and asks which service and which tier you would recommend.

Relational data

Which service

Service	Choose it when
Azure SQL Database	You build or modernise an application and want a managed single database or elastic pool
Azure SQL Managed Instance	You lift and shift SQL Server and need near-complete engine compatibility, such as SQL Agent or cross-database queries
SQL Server on an Azure VM	You need OS-level control or a feature the managed options lack
Managed PostgreSQL or MySQL	The application already targets an open-source engine

The pattern is always the same: take the most managed service that still meets the compatibility requirement. A requirement for OS access or an unsupported feature pushes you down to a VM; “minimal administration” pushes you up.

AZ-400 source control strategy explained

Fri, 09 Oct 2026 00:00:00 +0000

Design and implement a source control strategy is 10-15% of AZ-400. It tests whether you can choose a branching model, enforce it with branch policies and branch protection rules, and keep a Git repository healthy: large files, repository size, permissions, tags and recovering or removing data. The questions are scenarios with a team size, a release cadence and a problem.

Branching strategies

The outline names three: trunk-based, feature branch and release branch.

CCNA 200-301 switching and access explained

Fri, 09 Oct 2026 00:00:00 +0000

Switching and Network Access is also worth 25% of the v2.0 CCNA. It covers switch-to-switch and switch-to-router links, edge port configuration, CDP and LLDP, Layer 2 and Layer 3 troubleshooting with show commands, and Rapid PVST+. Questions mix configuration snippets with output, and the traps are small details: native VLANs, LACP modes and bridge priority arithmetic.

Infrastructure connectivity

Topic	Key facts
802.1Q trunks	Tag every frame except the native VLAN; both ends must agree on the native VLAN and the allowed list
EtherChannel	LACP modes active and passive; two passive ends never form a bundle; member ports need matching speed, duplex and VLAN or trunk settings
SVI	A virtual interface per VLAN; used for management and for inter-VLAN routing on a Layer 3 switch
Routed ports	`no switchport` turns a physical port into a Layer 3 interface

Inter-VLAN routing on a Layer 3 switch needs ip routing plus an up/up SVI for each VLAN. Router-on-a-stick uses subinterfaces with encapsulation dot1q.

DP-600 prepare data explained

Fri, 09 Oct 2026 00:00:00 +0000

Prepare data is the largest DP-600 domain at 45–50%. It covers three skills: getting data into or connecting to it from Fabric, transforming it into an analysis-ready shape, and querying it with the visual query editor, SQL, KQL and DAX. If you only have time to go deep in one place, go deep here and practise in a real workspace.

Get data

Choosing a data store

The first decision in many scenarios is which item holds the data.

DP-700 ingest and transform data explained

Fri, 09 Oct 2026 00:00:00 +0000

Ingest and transform data is worth 30–35% of DP-700. It covers three areas: loading patterns, batch ingestion and transformation, and streaming. The exam asks you to choose an approach and to read or complete SQL, PySpark and KQL, so you need both judgement and syntax.

Loading patterns

Full versus incremental. A full load replaces everything and is simple, but its cost grows with the table. An incremental load moves only new or changed rows, usually using a watermark (a stored value such as the highest modified date or key already loaded) or change data capture. Choose full for small reference data; incremental when a large table changes slowly.

DP-750 Unity Catalog security explained

Fri, 09 Oct 2026 00:00:00 +0000

Secure and govern Unity Catalog objects is worth 15–20% of DP-750. It tests whether you can give the right people the right access and no more, authenticate pipelines without hard-coded secrets, and show where data came from and who touched it. Most questions describe a symptom (a user gets a permission error) or a requirement (hide a column from one group) and ask which Unity Catalog feature fits.

Secure Unity Catalog objects

Grant privileges

Privileges are granted to a principal (user, service principal or group) on a securable object and are inherited downward: a grant on a catalog applies to its schemas and tables. Granting to groups rather than individuals is the standard recommendation.

DP-800 secure, optimize and deploy

Fri, 09 Oct 2026 00:00:00 +0000

Secure, optimize, and deploy database solutions is the other 35–40% domain in DP-800, and it is the widest. It covers data security and compliance, performance tuning, CI/CD with SQL Database Projects, and integration with Azure services such as Data API builder. Questions usually describe a requirement and ask which feature meets it, or show a symptom and ask for the cause.

Implement data security and compliance

Feature	What it does	What it does not do
Always Encrypted	Encrypts columns client-side; the server never sees plaintext	Does not allow all operations on encrypted data; deterministic encryption allows equality, randomized does not
Column-level encryption	Encrypts values with keys managed in the database	Keys live on the server side, so it protects less from administrators
Dynamic Data Masking	Masks values in query results for users without `UNMASK`	Is not encryption; masked columns can be inferred by filtering
Row-Level Security	Filters rows, and blocks writes, by predicate	Does not hide columns
Object permissions	`GRANT`, `DENY`, schemas and roles control access	Cannot restrict individual rows

Row-Level Security uses an inline table-valued function as a predicate and a security policy to apply it. A filter predicate affects reads, updates and deletes; a block predicate stops writes that violate it. Use SESSION_CONTEXT when many users share one application login.

Enrolled Agent for accountants and tax preparers

Fri, 09 Oct 2026 00:00:00 +0000

An accountant or tax preparer should become an Enrolled Agent if the work is federal tax and the goal is unlimited representation rights before the IRS without a degree or a state licence. The credential needs all three parts of the Special Enrollment Examination: Part 1 covers individuals, Part 2 covers businesses, and Part 3 covers representation, practices and procedures. A CPA already has the same representation rights, so the EA adds nothing there; if the goal is audit or attest work, the CPA is the route.

PL-300 model the data explained

Fri, 09 Oct 2026 00:00:00 +0000

Model the data is worth 25–30% of PL-300 and is where most candidates lose the most marks. It covers designing the data model, writing calculations in DAX, and making the model fast. The questions describe tables and a requirement and ask which relationship, which expression or which setting satisfies it, so you need to reason about filter context rather than memorise syntax.

Design and implement a data model

Star schema and relationships

A good model is a star: one central fact table of events and dimension tables around it, joined by one-to-many relationships from dimension to fact. Filters flow from the one side to the many side by default, which is why a slicer on a dimension filters the fact table.

SC-200 respond to incidents explained

Fri, 09 Oct 2026 00:00:00 +0000

Respond to security incidents is worth 35–40% of SC-200 and tests whether you can work an incident from the first alert to remediation. It has three groups of objectives: responding in Defender XDR and the products feeding it, responding on devices with Defender for Endpoint, and investigating Microsoft 365 activity. Questions give you a situation and ask for the next action or the right tool.

Respond in Defender XDR

Defender XDR correlates alerts from several products into one incident, so the first skill is reading the incident: the attack story, the affected entities and the evidence. The outline lists the sources whose alerts you must know how to investigate and remediate:

SC-300 authentication and access explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement authentication and access management is the largest SC-300 domain at 25–30%. It covers how users prove who they are (authentication methods, MFA, SSPR, passwordless), when access is allowed (Conditional Access), how risk is handled (ID Protection) and how network access is secured (Global Secure Access). Expect scenario questions where a policy does not behave as expected.

User authentication

Methods

Authentication methods are enabled in the authentication methods policy, per method and per group. The ones the outline names:

SC-401 data loss prevention explained

Fri, 09 Oct 2026 00:00:00 +0000

Data loss prevention is the first half of the 30–35% domain Implement data loss prevention and retention. A Purview DLP policy decides where to watch, what counts as sensitive, and what to do when a match is found. The exam asks you to design a policy for a requirement, to predict what happens when several policies or rules apply, and to configure Endpoint DLP on devices.

Building a DLP policy

A policy has four moving parts:

SC-900 Microsoft Entra explained

Fri, 09 Oct 2026 00:00:00 +0000

The Microsoft Entra skill area is worth 25–30% of SC-900, the second largest. It covers what Microsoft Entra ID is and the identity types it holds, how users authenticate (methods, MFA, password protection), how access is controlled (Conditional Access, roles and RBAC), and how identities are protected and governed (ID Protection, Privileged Identity Management, access reviews and ID Governance). Most questions ask which feature meets a given requirement.

Microsoft Entra ID and identity types

Microsoft Entra ID is Microsoft’s cloud identity and access management service, used to sign in to Microsoft 365, Azure and thousands of other apps. It holds several identity types: users, groups, devices and workload identities (identities for applications and services rather than people, such as service principals and managed identities). External identities allow partners and customers to access resources.

SY0-801 threats, vulnerabilities and attacks explained

Fri, 09 Oct 2026 00:00:00 +0000

Threats, Vulnerabilities, and Attacks is worth 24% of SY0-801, the second-largest domain. It has six objectives (2.1 to 2.6) and mostly asks you to recognise something from a short scenario: who the attacker is, how they arrived, what weakness they used, and what evidence shows it. The newest objective, 2.6, covers threats linked to artificial intelligence.

2.1 and 2.2 Threats, vulnerabilities and actors

Know threat feeds, likelihood and impact, intelligence sources and the threat life cycle. For vulnerabilities, learn CVSS (a severity score used to prioritise) and CVE (an identifier for a specific known flaw). They are often confused.

AZ-802 hybrid management explained

Tue, 29 Sep 2026 00:00:00 +0000

Manage Windows Server instances and workloads in a hybrid environment is worth 10–15% of AZ-802. It has two halves: remote management of individual servers with Windows Admin Center, PowerShell, SSH and Remote Desktop, and management at scale through Azure with Azure Arc, machine configuration, VM extensions, Azure Update Manager and Azure Automation. Most questions ask which tool fits a requirement, or why a remote command fails.

Remote management

Windows Admin Center

Windows Admin Center is a browser-based management tool for servers, clusters and Hyper-V hosts. It runs in two ways:

AB-900 data protection and governance

Mon, 28 Sep 2026 00:00:00 +0000

Understand data protection and governance tasks for Microsoft 365 and Copilot is worth 35–40% of AB-900, the largest share of the exam. It covers four things: the Microsoft Purview solutions, how Copilot accesses data, the tools that surface data risks, and oversharing in SharePoint. The skill being tested is matching a problem to the right tool — and there are many tools that sound alike.

The Purview solutions

Solution	Answers the question
Information Protection	How do we classify and protect this content, wherever it goes?
Data Loss Prevention (DLP)	How do we stop sensitive content leaving where it should stay?
Data Lifecycle Management	How long do we keep this, and when do we delete it?
Insider Risk Management	Is someone inside behaving in a way that puts data at risk?
Communication Compliance	Are messages breaking our conduct or regulatory policies?
DSPM for AI	What are people doing with AI, and what sensitive data is involved?

Sensitivity labels

A sensitivity label classifies content and can protect it: encryption, content marking such as headers and watermarks, and settings for containers such as sites, groups and teams. Because encryption travels with the file, protection holds after the file is downloaded or emailed. Labels matter for Copilot too: Copilot respects the usage rights encryption grants, and content it creates from labelled sources can inherit the label.

AI-300 ML model lifecycle explained

Mon, 28 Sep 2026 00:00:00 +0000

Implement machine learning model lifecycle and operations is the largest AI-300 domain at 25–30%. It follows one model from training through registration and deployment to monitoring in production. Much of it will be familiar from DP-100, but the emphasis is on the operational end: how a model is versioned, rolled out safely, rolled back, and retrained when the data changes.

Orchestrating training

Jobs and MLflow

A command job runs a script on a compute target in an environment. Azure Machine Learning workspaces are MLflow tracking servers, so mlflow.autolog() or explicit mlflow.log_metric calls record parameters, metrics and artifacts without extra configuration inside a job. Comparing runs in the studio, or querying them with the MLflow client, is how you compare model performance across jobs.

AZ-104 storage explained

Mon, 28 Sep 2026 00:00:00 +0000

Implement and manage storage is worth 15–20% of AZ-104. It covers three areas: controlling access to storage accounts, creating and configuring the accounts themselves, and working with Azure Files and Blob Storage. The questions are detailed. You need to know which redundancy option survives which failure, which kind of SAS can be revoked, and which blob feature needs another feature switched on first.

Configure access to storage

Access keys

Every storage account has two access keys. Two, so you can rotate without downtime: move applications to key 2, regenerate key 1, then repeat later. A key grants full access to the whole account, which is why most scenarios steer you away from sharing it.

AZ-900 architecture and services explained

Mon, 28 Sep 2026 00:00:00 +0000

Describe Azure architecture and services is the largest AZ-900 domain at 35–40% of the exam. It covers four areas: the core architectural components of Azure, compute and networking, storage, and identity, access and security. It is also the broadest domain, and most wrong answers come from confusing two similar services. Learn each service together with the one it is most often mistaken for.

Core architectural components

Physical structure

Component	What it is	Protects against
Datacenter	A facility full of servers	—
Availability zone	One or more datacenters in a region with independent power, cooling and networking	Loss of a datacenter
Region	A geographic area with one or more datacenters close together	—
Region pair	Two regions in the same geography, paired for recovery and staggered updates	Loss of a whole region
Sovereign region	An isolated instance of Azure for government or legal requirements	Compliance, not outages

Not every region has availability zones. Services that use zones are either zonal (pinned to one zone) or zone-redundant (replicated across zones).

AI-200 data management services explained

Sun, 27 Sep 2026 00:00:00 +0000

Develop AI solutions by using Azure data management services is the largest AI-200 domain at 25–30%, and the newest. It covers three services: Azure Cosmos DB for NoSQL, Azure Database for PostgreSQL and Azure Managed Redis. Each stores embeddings and runs vector similarity search, so the exam tests both how each service works and which one fits a scenario.

Choosing between the three

Service	Reach for it when
Cosmos DB for NoSQL	Documents with flexible schema, global scale, change-driven processing
PostgreSQL with pgvector	Relational data with joins and constraints, SQL-based RAG with metadata filters
Azure Managed Redis	Sub-millisecond reads, caching, short-lived or frequently repeated lookups

The recurring trap is picking Redis as a primary store. It is a cache and a fast index. When a scenario needs durable system-of-record data, the answer is one of the other two.

SEE Part 1 income and assets explained

Mon, 21 Sep 2026 00:00:00 +0000

Income and Assets is worth 17 of the 85 scored questions on SEE Part 1, 20% of the paper. Together with Deductions and Credits it is one of the two largest domains. It asks what counts as income and when, how retirement distributions are taxed, how to find the basis and character of property that is sold, and which adjustments reduce gross income. It is also the widest domain in the outline, so you need to cover a lot of ground.

SEE Part 2 business tax preparation explained

Mon, 21 Sep 2026 00:00:00 +0000

Business Tax Preparation is the largest domain on SEE Part 2: 37 of the 85 scored questions, about 44%. It covers the working content of a business return — income, cost of goods sold, deductions and credits, depreciation, asset basis and dispositions, reading financial records, reconciling books to tax — plus advising the business client on filing, deposits, records, entity choice and worker classification. Every dollar figure below is a 2025 figure checked against the 2025 IRS instructions or publications.

SEE Part 3 representation explained

Mon, 21 Sep 2026 00:00:00 +0000

Representation before the IRS is worth 25 of the 85 scored questions on SEE Part 3 (about 29%). It covers the paperwork that lets you act for a client — Form 2848, Form 8821 and the CAF system — and the preliminary work of building a case: identifying the issue, assessing finances, finding authority and watching deadlines.

Form 2848 versus Form 8821

The single most tested distinction in this domain.

AAIA AI operations explained

Sun, 20 Sep 2026 00:00:00 +0000

AI Operations is worth 46% of AAIA — around 41 of the 90 questions, nearly half the exam and more than the other two domains combined. It is also the domain with the least audit method in it.

That combination is why candidates fail. An experienced auditor assumes an audit credential will reward audit knowledge, allocates study time accordingly, and meets a paper where almost half the questions are about how AI systems are built, changed, supervised and broken.

AAIR AI life cycle risk management

Sun, 20 Sep 2026 00:00:00 +0000

AI Life Cycle Risk Management is worth 21% of AAIR — around 19 of the 90 questions and the smallest of the three domains. It sounds like the most technical part of the exam, which is why candidates over-prepare it and under-prepare the two risk domains that together make up 79%.

You need to know what happens at each stage and what risk it introduces. You do not need to be able to perform any of it.

AAISM AI risk management explained

Sun, 20 Sep 2026 00:00:00 +0000

AI Risk Management is worth 31% of AAISM — around 28 questions. ISACA describes it as assessing and managing risks, threats, vulnerabilities and supply chain issues related to enterprise-wide AI adoption.

The methodology here is the one you already use. What changes is the catalogue: new sources of risk, new threats that target stages traditional systems do not have, and a supply chain that includes things you cannot inspect.

Risk assessment, thresholds and treatment

Standard discipline applied to a non-standard asset. Three AI-specific wrinkles are worth holding:

AB-620 integrate and extend agents explained

Sun, 20 Sep 2026 00:00:00 +0000

Integrate and extend agents in Copilot Studio is worth 40–45% of AB-620 — the largest domain and the one that distinguishes this exam from every other agent certification. It is a catalogue of mechanisms for connecting an agent to the outside world, and the exam’s favourite question shape is: here is a requirement, which mechanism fits?

Learn it as a decision table, not as a list.

The decision table

Requirement	Mechanism
Reach a common SaaS system with a ready-made integration	A Copilot connector or Power Platform connector
Reach an internal service that has a documented API	A REST API tool, or a custom connector
Expose tools to the agent through an open, model-facing standard	MCP tools
Search your own indexed content for grounding	Azure AI Search
Reach a system with no API at all	Computer use
Query enterprise data in Fabric	A Fabric data agent
Hand work to another agent on another platform	The A2A protocol
Hand work to a specialist agent you built in Foundry	Foundry agent integration

If you can reproduce that table from memory, you have most of this domain.

Agentforce Specialist Data 360 explained

Sun, 20 Sep 2026 00:00:00 +0000

Data 360 Fundamentals is worth 20% of the Salesforce Agentforce Specialist exam — around 12 of the 60 scored questions. It is the shortest section in the outline and the one candidates most often treat as background reading.

It is not background. It is the machinery that makes grounding work, and when an agent gives a confidently wrong answer, this is almost always where the fault lies.

Salesforce names just two objectives, and both matter.

AI-103 generative AI and agentic solutions explained

Sun, 20 Sep 2026 00:00:00 +0000

Implement generative AI and agentic solutions is the largest domain on AI-103 at 30–35% of the score, and it is the reason the exam exists. Where the older track spread attention across individual Azure AI services, this domain concentrates on one thing: building an application or an agent on Microsoft Foundry that uses a model, grounds its answers in your data, calls tools, and keeps working once real users touch it.

AI-901 implement AI solutions with Foundry explained

Sun, 20 Sep 2026 00:00:00 +0000

Implement AI solutions by using Microsoft Foundry is worth 55–60% of AI-901 — more than the concepts domain and the single biggest reason people who studied for AI-900 fail this exam. It is hands-on. Microsoft’s audience profile assumes Python syntax and familiarity with REST APIs, SDKs and CLIs, and the objectives are written as things you do, not things you describe.

You will not be asked to write a full application. You will be asked which portal step, which SDK call, or which service produces a stated result — and those questions are much easier if you have actually done it once.

AIF-C01 fundamentals of GenAI explained

Sun, 20 Sep 2026 00:00:00 +0000

Fundamentals of GenAI is worth 24% of AIF-C01’s scored content — around 12 of the 50 scored questions. Together with applications of foundation models it makes generative AI 52% of the exam, which is the single most important planning fact about AIF-C01.

Foundation models

A foundation model is a large model trained on a broad dataset, adaptable to many tasks without being built for any one of them. That adaptability is the point: rather than training a separate model per task, you take a general one and steer it.

CY0-001 securing AI systems explained

Sun, 20 Sep 2026 00:00:00 +0000

Securing AI systems is worth 40% of CY0-001 — roughly twenty-four of the sixty questions, and more than double any other domain. If you pass or fail this exam, this is usually where it happens. CompTIA describes it as implementing security controls for AI systems, protecting training data, models and outputs, and securing deployment environments across cloud and on-premises.

The mental model that makes it manageable: an AI system has four assets, and each needs its own controls.

Databricks Data Analyst: executing queries

Sun, 20 Sep 2026 00:00:00 +0000

Executing queries using Databricks SQL and Databricks SQL Warehouses is worth 20% of the Data Analyst Associate exam — around 9 of the 45 scored questions and the largest section. It covers writing SQL that is correct, and running it on compute that is appropriate.

The SQL you are expected to write

Standard analytical SQL, at a level a working analyst should have:

Joins — inner, left, and the consequences of each. Knowing why a join can multiply rows.
Aggregations with GROUP BY and HAVING, and the difference between filtering before and after aggregation.
Subqueries and CTEs — and preferring CTEs for readability in multi-step logic.
Window functions — ranking, running totals, lag and lead, period-over-period comparison. The exam guide does not name them, but they are everyday analyst SQL and make multi-step query scenarios easier to reason about.
Date and string handling.

The most examined correctness trap is the join that inflates row counts. If a fact table has 50,000 rows and joining a dimension produces 73,000, the join key is not unique on the dimension side. Analysts meet this constantly in practice and the exam knows it.

Databricks Data Engineer: transformation and modelling

Sun, 20 Sep 2026 00:00:00 +0000

Data Transformation and Modeling is worth 22% of the Databricks Data Engineer Associate exam — around 10 of the 45 scored questions and the largest single section. It covers turning raw ingested data into something usable, and doing it in a way that survives being rerun.

Delta Lake operations

The mechanics you are expected to know:

Operation	What it does
MERGE	Insert, update or delete in one statement based on a match condition — the upsert
UPDATE / DELETE	Modify rows in place, which plain file formats cannot do
Time travel	Read a previous version of a table
OPTIMIZE	Compact small files into larger ones
VACUUM	Remove files no longer referenced, subject to a retention window

MERGE is the one to know cold. It is how you apply changes without duplicating, and it is the standard answer to “records are being double-counted on rerun”.

Databricks DE Professional: cost and performance

Sun, 20 Sep 2026 00:00:00 +0000

Cost & Performance Optimisation is Section 5 of the Databricks Data Engineer Professional exam and worth 15% in the outline that applies from 9 October 2026: around 9 of the 60 scored questions and the second largest section. It grew from 13% and took over the Liquid Clustering versus partitioning comparison that used to sit under data modelling. It has no equivalent at Associate level, where tuning is folded into a general troubleshooting section.

Databricks GenAI application development

Sun, 20 Sep 2026 00:00:00 +0000

Application Development is worth 30% of the Databricks Generative AI Engineer Associate exam — the largest section, around 14 of the 45 scored questions. This is where you build the thing: the chain, the retrieval, the prompts and the behaviour when something goes wrong.

The RAG pattern, end to end

Almost everything in this section is a variation on one pipeline:

A user asks a question.
The question is embedded and matched against the index.
The top matching chunks are retrieved.
Those chunks are inserted into a prompt alongside the question and instructions.
The model answers from that context.
The response is returned, often with sources.

Know each step well enough to say what goes wrong when it fails, because that is how questions are framed.

Databricks ML Associate model development

Sun, 20 Sep 2026 00:00:00 +0000

Model Development is worth 31% of the Databricks Machine Learning Associate exam — around 15 of the 48 scored questions and the second largest section. This is the modelling work itself, framed as things you do on Databricks in Python.

Preparing data for modelling

Exploratory analysis: distributions, missing values, outliers, class balance.
Handling missing data, and recognising that each choice is a modelling decision.
Encoding categorical variables; scaling numeric ones where the algorithm needs it.
Splitting into training, validation and test sets appropriately for the data — and chronologically for time series, never randomly.

Feature engineering

Creating features that carry signal.
Reusing feature definitions between training and inference so they are computed identically. Divergence here is training/serving skew, a classic production failure.

Training

Common algorithm families and the problems they suit: linear models, tree ensembles, gradient boosting.
Training in Python on Databricks, and when distributed training is warranted.
Managing compute cost during training.

Hyperparameter tuning

Hyperparameters are chosen before training — learning rate, tree depth, regularisation, number of estimators. Distinguish them from parameters, which the model learns.
Search strategies, and automated tuning at scale.
Tuning against a validation set, never the test set.

Evaluation metrics

The most reliably tested area in the section, because metric choice is where people go wrong.

Databricks ML Professional: ML Ops

Sun, 20 Sep 2026 00:00:00 +0000

ML Ops is worth 44% of the Databricks Machine Learning Professional exam — around 26 of the 59 scored questions, tied with model development as the largest section.

It has no counterpart on the ML Associate, and it is the single clearest reason the Professional expects a year of experience rather than six months. Most of this material describes problems that only appear after a model has been live for months.

The founding idea

A machine learning system degrades on its own.

MLA-C02 model and FM development

Sun, 20 Sep 2026 00:00:00 +0000

ML Model and Foundation Model (FM) Development is worth 24% of MLA-C02. AWS frames it as selecting modelling approaches, training models, tuning hyperparameters, analysing performance and managing model versions — now covering both traditional ML and foundation models.

Choosing an approach

Given a business problem, pick the modelling approach:

Classification for categories, regression for continuous values, clustering for unlabelled grouping, forecasting for time series.
Built-in algorithms versus your own code versus a pre-trained service.
And the question C02 adds: traditional ML or a foundation model?

That last decision has a practical answer. A well-defined, high-volume, narrow task — classifying transactions, predicting a number — is usually cheaper and more reliable with a traditional model. An open-ended language or multimodal task is foundation model territory.

N10-009 network implementation explained

Sun, 20 Sep 2026 00:00:00 +0000

Network implementation is worth 20% of CompTIA Network+ N10-009 — around 18 of the 90 questions. This is the building domain: routing, switching, wireless and physical installation.

Routing

Static versus dynamic routing, and when each is appropriate. Static suits small, stable networks; dynamic scales and adapts.
Routing protocol families at a conceptual level: distance vector, link state, and the distinction between interior and exterior protocols.
The default route and what it does.
How a router selects a path when several exist, and what a routing table contains.
Router-on-a-stick and inter-VLAN routing — how VLANs talk to each other.

Route selection: prefix length, administrative distance and metric

Objective 2.1 names three route selection criteria: administrative distance, prefix length and metric. Objective 5.3 in the troubleshooting domain comes back to the same ground with the routing table and default routes. Network+ wants you to apply them in the right order, not to memorise every protocol’s internals.

Salesforce AI certification path

Sun, 20 Sep 2026 00:00:00 +0000

The Agentforce Specialist exam assumes you already know Salesforce. It is not an AI exam that happens to run on Salesforce; it is a Salesforce exam about building agents, and the platform knowledge is the harder prerequisite for most candidates.

What the exam actually weights

Section	Weight
AI Agents	35%
Prompt Engineering	20%
Data 360 Fundamentals	20%
Testing, Deployment, and Maintenance	10%
Governance and Observability	10%
Multi-Agent Orchestration	5%

Over a third of the score is agent construction: topics, actions, scope, escalation and how the security model constrains what an agent can reach. Another fifth is data — identity resolution, harmonisation, ingestion — which is where people from an AI background usually lose marks.

SC-500 storage, databases and networking

Sun, 20 Sep 2026 00:00:00 +0000

Secure storage, databases, and networking is worth 25–30% of SC-500 — the largest domain, though only just. It is also the most conventional: this is classic Azure security engineering, largely unchanged from AZ-500 apart from two newer networking objectives.

Storage accounts

Four objectives: security configuration, storage firewall rules, Defender for Storage threat protection, and managing access including access policies.

The pattern the exam uses is layering. A storage account is not secured by one control:

Spark Developer: the DataFrame API

Sun, 20 Sep 2026 00:00:00 +0000

Developing Apache Spark DataFrame/DataSet API Applications is worth 30% of the Spark Developer Associate exam — around 14 of the 45 scored questions and the largest section. All code is Python.

The exam is multiple choice, so you read PySpark rather than write it. But the fluency required is precise: you will be shown a snippet and asked what it returns, which of four versions is correct, or what is wrong with it.

200-901 Cisco platforms explained

Fri, 09 Oct 2026 00:00:00 +0000

Cisco Platforms and Development is 15% of 200-901. It is the part of the exam that is specific to Cisco, and the part with the most facts to memorise. You need to know what each platform manages, how its API authenticates, what the typical objects are, and how to read or write a short script that calls it. Cisco names the platforms in the exam topics, so the list below follows that outline.

AB-100 Deploy AI solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

Deploy AI-powered business solutions is the largest AB-100 domain at 40–45%, and it is mostly about running AI responsibly after it is built. It has four parts: analysing, monitoring and tuning; managing testing; designing the ALM process; and designing responsible AI, security, governance, risk management and compliance. Questions ask what to measure, what to gate, how to promote changes and how to protect data.

Analyse, monitor and tune

The outline asks you to recommend the process and tools for monitoring agents, analyse the backlog and user feedback, use AI-based tools to find issues and tune, monitor performance and metrics, and interpret telemetry for performance and model tuning.

AB-730 business content and collaboration explained

Fri, 09 Oct 2026 00:00:00 +0000

Manage business content and collaboration by using Microsoft 365 Copilot is worth 20–25% of AB-730. It tests how you use Copilot to draft and analyse content from the right sources across Word, Excel and PowerPoint, how you check the result against the business goal, and how you use the AI features around meetings and shared work. The skill underneath is judgement: choosing the right source and the right app, then verifying.

AB-731 implementation and adoption explained

Fri, 09 Oct 2026 00:00:00 +0000

Identify an implementation and adoption strategy for Microsoft’s AI apps and services is worth 20–25% of AB-731. It covers two things: aligning AI use with responsible AI and governance, and planning adoption across the organisation. It is the smallest domain, and the one that most resembles the real job of a transformation leader.

Responsible AI and governance

The outline lists the standards AI solutions must meet. Learn each as a one-line test:

AZ-305 business continuity explained

Fri, 09 Oct 2026 00:00:00 +0000

Design business continuity solutions is worth 15–20% of AZ-305. It has two halves: backup and disaster recovery, and high availability. Almost every question can be solved by reading two numbers from the scenario, the recovery point objective (RPO, how much data you can lose) and the recovery time objective (RTO, how long you can be down), and then choosing the cheapest design that meets both.

Start with the objectives

Term	Question it answers
RPO	How much recent data can we afford to lose?
RTO	How long can the service be unavailable?
Availability target	What percentage of the time must the service be up?

Tighter objectives cost more. A request for an RPO of seconds rules out daily backups; an RTO of minutes rules out restoring from backup into an empty region. Recommend per tier: a reporting database seldom needs the same design as the order system.

AZ-400 build and release pipelines explained

Fri, 09 Oct 2026 00:00:00 +0000

Design and implement build and release pipelines is 50-55% of AZ-400, more than all other domains together. It covers six areas: package management, testing in pipelines, pipeline design, deployment strategies, infrastructure as code, and pipeline maintenance. If you can read and write a multi-stage YAML pipeline and justify a deployment strategy from a requirement, you have most of the exam.

Package management

Tools. Azure Artifacts and GitHub Packages host packages (NuGet, npm, Maven, Python, containers and more).
Feeds and views. An Azure Artifacts feed can pull from upstream sources such as nuget.org, and views (for example @Local, @Prerelease, @Release) promote a package through quality levels without republishing it.
Versioning. Semantic versioning (MAJOR.MINOR.PATCH) signals breaking changes; calendar versioning (CalVer) uses dates. Pipeline artifacts also need a versioning strategy, often the build number.

Testing strategy for pipelines

Tests are layered: local tests, unit tests, integration tests and load tests, each running where it is cheap and meaningful. You should know how to publish test results, collect code coverage, and add quality gates and release gates (a failed coverage threshold, a security scan result, an Azure Monitor alert check or a manual approval can block a release). Flaky tests, which pass and fail without code change, are a maintenance topic: quarantine them rather than ignore the pipeline.

CCNA 200-301 IP routing explained

Fri, 09 Oct 2026 00:00:00 +0000

IP Routing is 20% of the v2.0 CCNA. It covers reading the routing table to find the next hop, troubleshooting IPv4 and IPv6 static routes, configuring single-area OSPFv2 and OSPFv3, and interpreting the state of first hop redundancy protocols. It rewards a precise mental model: how a router chooses among routes, and what has to match before OSPF routers become neighbors.

How a router picks a route

Three rules, applied in order:

Data analyst moving to Fabric or Databricks

Fri, 09 Oct 2026 00:00:00 +0000

A data analyst moving to Microsoft Fabric should take PL-300 if reports and semantic models are the daily work and then DP-600 once the job includes lakehouses, warehouses and SQL; an analyst moving to Databricks should take the Databricks Data Analyst Associate. The engineering exams (DP-700, DP-750, Databricks Data Engineer Associate) are the right answer only if the next job is building pipelines rather than querying and presenting data.

DP-600 semantic models explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement and manage semantic models is worth 25–30% of DP-600. It splits into designing and building models (storage modes, schema, relationships, DAX, calculation groups, composite models) and optimising them at enterprise scale (performance, Direct Lake, incremental refresh). Scenario questions here usually ask which choice avoids a problem, such as slow queries, fallback or a refresh that takes too long.

Design and build semantic models

Storage modes

Mode	How data is read	Typical trade-off
Import	Data is loaded into the model’s in-memory store on refresh	Fast queries; data is as fresh as the last refresh
DirectQuery	Queries go to the source at query time	Fresh data; speed depends on the source
Dual	Table can behave as Import or DirectQuery depending on the query	Useful for dimensions in composite models
Direct Lake	Reads Delta tables in OneLake directly	Import-like speed without a full data copy

Choose by freshness, volume and speed. A composite model mixes modes across tables, for example Import dimensions with a DirectQuery fact table, and can also combine several sources. The large semantic model storage format lets models grow past the default size limit and is the option to know for large models.

DP-700 monitor and optimize explained

Fri, 09 Oct 2026 00:00:00 +0000

Monitor and optimize an analytics solution is worth 30–35% of DP-700, the same as each other domain. It covers monitoring Fabric items, identifying and resolving errors, and optimizing performance. Many candidates under-prepare it because it feels like operations rather than engineering, yet it is a third of the marks.

Monitor Fabric items

The Monitor hub lists activity across items you can access: pipeline runs, Dataflow Gen2 refreshes, notebook and Spark job runs, and semantic model refreshes. From a run you can drill into details such as duration, status and error messages. For Spark, the run details link to the Spark UI and logs.

DP-750 prepare and process data explained

Fri, 09 Oct 2026 00:00:00 +0000

Prepare and process data is the largest DP-750 domain, at 30–35%. It covers four jobs: designing how data is modelled in Unity Catalog, ingesting it, cleansing and transforming it, and enforcing quality. Expect many short scenarios where one detail (batch or streaming, history or current state, upsert or append) decides the answer.

Design and implement data modeling

Formats, loading and tables

Choose a table format from Parquet, Delta, CSV, JSON or Iceberg. For lakehouse tables Delta is the default because it adds transactions, schema enforcement, time travel and MERGE. CSV and JSON are source formats you ingest from, not formats you usually keep. Choose batch loading for periodic, bounded data and streaming when data arrives continuously and latency matters. Choose managed tables unless files must live somewhere you control.

DP-800 AI capabilities explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement AI capabilities in database solutions is 25–30% of DP-800. It covers three areas: external models and embeddings, intelligent search (full-text, vector and hybrid) and retrieval-augmented generation (RAG) driven from T-SQL. The questions are practical: which search to choose, which embedding maintenance method fits, and how a prompt is assembled from query results.

Design and implement models and embeddings

An embedding is a vector of numbers produced by a model so that texts with similar meaning end up close together. The database stores those vectors and compares them.

PL-300 visualize and analyze explained

Fri, 09 Oct 2026 00:00:00 +0000

Visualize and analyze the data is worth 25–30% of PL-300. It covers building reports with the right visuals, making them usable and tell a story, and using Power BI’s analytical features to find patterns. This is the area most analysts already know, so the exam mixes familiar tasks with items people skip: accessibility, mobile layouts, paginated reports, and the newer Copilot and visual calculation features that the current outline names.

SC-200 threat hunting explained

Fri, 09 Oct 2026 00:00:00 +0000

Perform threat hunting is worth 20–25% of SC-200 and is about looking for threats that no detection has caught yet. It is the smallest domain, but it asks the most of your KQL, and the same skills sit behind custom detections and analytics rules in the first domain. The outline splits it into hunting with Microsoft Defender XDR and hunting on the Microsoft Sentinel platform.

Detect threats with Defender XDR

The objectives are to identify the right table for a KQL query, identify threats with KQL, create advanced hunting queries, interpret threat analytics, create hunting graphs including blast radius, and analyze relationships between entities with Sentinel Graph.

SC-300 workload identities explained

Fri, 09 Oct 2026 00:00:00 +0000

Plan and implement workload identities is worth 20–25% of SC-300. It covers identities for applications and Azure workloads (managed identities and service principals), enterprise applications, app registrations and app access monitoring with Microsoft Defender for Cloud Apps. The main skill is choosing the right identity or permission type for a described application.

Identities for applications and Azure workloads

Identity	Use when
System-assigned managed identity	One Azure resource needs credentials; the identity is deleted with the resource
User-assigned managed identity	Several resources share an identity, or it must outlive a resource
Service principal (with secret, certificate or federated credential)	The workload runs outside Azure, or in a pipeline, and needs its own identity
User account	Interactive or legacy scenarios that cannot use a workload identity (avoid for automation)
Managed service account	On-premises Windows services running under AD

The preferred answer in nearly every scenario is a managed identity, because there is no secret to store or rotate. You create one, assign it to an Azure resource, give it an RBAC role on the target resource, and the app requests tokens without credentials.

SC-401 retention explained

Fri, 09 Oct 2026 00:00:00 +0000

Retention is the second half of the 30–35% domain Implement data loss prevention and retention. Retention in Microsoft Purview means keeping content for a period, deleting it after a period, or both. The exam tests the choice between retention policies and retention labels, how scopes are targeted, what happens when several retention rules overlap, and how to get content back.

Policy or label?

	Retention policy	Retention label
Applies to	Whole locations: mailboxes, sites, Teams, groups	Individual items
Targeting	Locations, optionally narrowed by scope	Published for users to choose, or auto-applied by condition
Can mark as record	No	Yes, including regulatory records
Disposition review	No	Yes
Typical use	Delete all Teams chats after two years	Keep each contract seven years, then review

Use a policy when the rule is the same for everything in a location. Use a label when different items need different treatment, or when you need records, event-based retention or a disposition review.

SC-900 Microsoft security solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft security solutions is the largest SC-900 skill area at 35–40%. It has four parts: core Azure infrastructure security services (DDoS Protection, Firewall, WAF, network segmentation, NSGs, Bastion, Key Vault), security management with Microsoft Defender for Cloud, Microsoft Sentinel as SIEM and SOAR, and the Microsoft Defender XDR family. The challenge is not depth but telling similar products apart, so learn each by the one problem it solves.

Core infrastructure security in Azure

Service	What it does
Azure DDoS Protection	Defends public IP resources against volumetric and protocol denial-of-service attacks
Azure Firewall	Managed, stateful network firewall with application, network and threat-intelligence rules
Web Application Firewall (WAF)	Protects web apps from common attacks such as SQL injection and cross-site scripting, on Application Gateway or Front Door
Virtual networks and segmentation	Split workloads into isolated subnets to limit lateral movement
Network security groups (NSGs)	Allow or deny traffic by source, destination, port and protocol at the subnet or NIC
Azure Bastion	Browser-based RDP and SSH to VMs over TLS, without public IP addresses on the VMs
Azure Key Vault	Stores and controls access to secrets, keys and certificates

An NSG is basic filtering; Azure Firewall is a full managed firewall; a WAF understands web traffic. If the scenario mentions SQL injection, the answer is WAF.

SY0-801 security architecture explained

Fri, 09 Oct 2026 00:00:00 +0000

Security Architecture is worth 19% of SY0-801. It has four objectives: comparing architecture models (3.1), protecting infrastructure (3.2), protecting data (3.3) and resilience and recovery (3.4). Questions ask you to choose a design or a data-protection method that fits a business requirement, so read for constraints such as cost, downtime tolerance and regulation.

3.1 Architecture models

The objective lists cloud, serverless, multicloud, hybrid, private, public and community deployment models, infrastructure as code, operational technology, on-premises, air-gapped networks, microservices, and logical versus physical segmentation. Know the trade-offs the objective names: availability, resilience, proprietary versus open source, usability, shared responsibility, compute and power requirements, ease of recovery, data sovereignty, data classification, cost, ownership, environmental requirements, scalability and risk.

AZ-802 virtual machines explained

Tue, 29 Sep 2026 00:00:00 +0000

Manage virtual machines is worth 10–15% of AZ-802. Most of it is Hyper-V: guest access, memory, device assignment, checkpoints, virtual disks and switches, NIC teaming, high availability and Hyper-V Replica. The rest covers Windows Server VMs in Azure: disks, sizing, availability and secure access. Questions tend to ask which setting enables a feature and what has to be true before you can turn it on.

Working with guests

Feature	What it gives you	Requirement
Enhanced Session Mode	Clipboard, drives, audio and USB redirection in VMConnect	Enabled on the host; guest supports it
PowerShell Direct	PowerShell into a guest without networking	Run on the host; guest credentials; `-VMName`
SSH Direct	SSH into a Linux guest over the VM bus, without networking	Supported Hyper-V host and Linux guest
Nested virtualization	Hyper-V inside a VM	VM off; `Set-VMProcessor -ExposeVirtualizationExtensions $true`

For nested VMs to reach the network, enable MAC address spoofing on the outer VM’s adapter or use NAT.

AB-900 Copilot and agent admin explained

Mon, 28 Sep 2026 00:00:00 +0000

Perform basic administrative tasks for Copilot and agents is worth 25–30% of AB-900. It covers three things: what Copilot and agents can do, the day-to-day administration of Copilot, and the lifecycle of agents from creation to monitoring. The skill being tested is operational: who gets what, who pays for it, and how you know whether it is being used.

Copilot and agents: the vocabulary

Term	Meaning
Microsoft 365 Copilot	The licensed assistant inside Word, Excel, PowerPoint, Outlook, Teams and the Copilot app, grounded in the user’s work data
Copilot Chat	The chat experience available to Microsoft 365 users; work-data grounding and agents can be added
Agent	A focused assistant with its own instructions, knowledge and actions, used inside Copilot
Researcher	A built-in reasoning agent for multi-step research across work data and the web, producing a structured report
Analyst	A built-in reasoning agent that works through data, such as spreadsheets, to find trends and produce analysis
Custom agent	An agent your organisation builds for a specific process, knowledge base or team

When to pick which

Researcher when the task is gathering and synthesising information: a market overview, a briefing before a customer meeting, a summary of everything written about a project.
Analyst when the task is working through numbers: finding patterns in sales data, forecasting, explaining why a metric moved.
A custom agent when the same kind of question keeps coming up over the same content: an HR policy agent, an IT help desk agent, a product specification agent grounded on one SharePoint site.

Licensing: monthly versus pay-as-you-go

	Per-user monthly licence	Pay-as-you-go
Who pays	Licence per named user	Metered by use, billed to an Azure subscription
Fits	People who use Copilot daily across apps	Occasional users, or agents for staff without a licence
Set up in	Microsoft 365 admin center, licences	Microsoft 365 admin center, billing policy linked to Azure
SharePoint agents	Included for licensed users	Metered for users without a Copilot licence

A billing policy connects pay-as-you-go usage to an Azure subscription and resource group, and scopes it to users or groups. Admins monitor spend there and can use Azure cost management budgets to cap surprises.

AI-300 GenAIOps infrastructure explained

Mon, 28 Sep 2026 00:00:00 +0000

Design and implement a GenAIOps infrastructure is worth 20–25% of AI-300. It is the generative AI counterpart of the MLOps infrastructure domain: set up Microsoft Foundry securely and from code, deploy foundation models in the way that fits the workload, and treat prompts as versioned source rather than text pasted into a portal.

Foundry resources and projects

A Foundry resource is the Azure resource that holds model deployments, connections and security settings. Projects live inside it and give teams a workspace for agents, evaluations and files, sharing the resource’s deployments and connections. A typical pattern is one resource per environment or business unit, with projects per application or team.

AZ-104 compute resources explained

Mon, 28 Sep 2026 00:00:00 +0000

Deploy and manage Azure compute resources is one of the two largest AZ-104 domains at 20–25%, and it has the most objectives. It covers four areas: automating deployments with ARM templates and Bicep, virtual machines and scale sets, containers in the portal, and Azure App Service. The questions range from reading a template to choosing between availability sets and zones.

ARM templates and Bicep

Reading a template

An ARM template is JSON with a fixed set of sections: parameters (values supplied at deployment), variables (values computed inside the template), resources (what gets deployed) and outputs (values returned afterwards). Bicep describes the same resources in a shorter syntax and compiles to ARM JSON.

AZ-900 management and governance explained

Mon, 28 Sep 2026 00:00:00 +0000

Describe Azure management and governance is the AZ-900 domain worth 30–35% of the exam. It answers four practical questions about running Azure: what it costs, how you keep it compliant, how you deploy and manage it, and how you know it is healthy. The difficulty is that many tools sound alike. Almost every question in this domain comes down to picking the right one of two or three similar tools.

AI-200 connect to Azure services explained

Sun, 27 Sep 2026 00:00:00 +0000

Connect to and consume Azure services is worth 20–25% of AI-200. It has two parts: event- and message-based solutions with Azure Service Bus and Azure Event Grid, and serverless code with Azure Functions. The skill being tested is moving slow work — embedding documents, calling models, writing to databases — off the request path, and handling it reliably when something fails.

Service Bus or Event Grid?

The first decision in almost every messaging question.

IRS enrolled agent certification path

Mon, 21 Sep 2026 00:00:00 +0000

There is one route to the enrolled agent credential: pass all three parts of the IRS Special Enrollment Examination, then apply for enrollment. The only real choice is the order. If you prepare individual returns or are new to tax, take Part 1 first, then Part 2, then Part 3. If you come from outside tax altogether, starting with Part 3 is a reasonable alternative.

The three exams

Exam	Content	Assumes	Largest domain
SEE Part 1	Individuals: Form 1040, income, deductions, credits, estate and gift tax	Familiarity with an individual return	Income and Assets / Deductions and Credits (17 each)
SEE Part 2	Businesses: sole proprietors, partnerships, C and S corporations, trusts, exempt entities	Part 1 knowledge, especially basis and pass-through income	Business Tax Preparation (37)
SEE Part 3	Representation, Practices and Procedures: Circular 230, powers of attorney, collection, examinations, e-file	Understanding of the returns being represented	Practices and Procedures (26)

All three share the same format: 100 questions of which 85 are scored, 3.5 hours, a scaled score from 200 to 800 with 500 to pass, and a $317 fee per part. There is no enforced order and the IRS lists no degree or experience requirement. You do need a PTIN before PSI will let you schedule.

SEE Part 1 deductions and credits explained

Mon, 21 Sep 2026 00:00:00 +0000

Deductions and Credits is worth 17 of the 85 scored questions on SEE Part 1, 20% of the paper, tied with Income and Assets as the largest domain. It covers the choice between the standard deduction and itemizing, each Schedule A category and its limits, the qualified business income deduction, and the individual credits: who qualifies, what they are worth, and whether they are refundable. Paid preparer due diligence for the earned income credit is also listed explicitly.

SEE Part 2 specialized returns explained

Mon, 21 Sep 2026 00:00:00 +0000

Specialized Returns and Taxpayers is the smallest SEE Part 2 domain — 18 of the 85 scored questions, about 21% — and the broadest. It covers five unrelated topics: trust and estate income tax, exempt organisations, retirement plans, farmers, and rental property. At three or four questions per topic you cannot skip any, but you do not need the depth the entity rules require: learn each topic’s core rules, filing requirements and signature traps. All dollar figures are 2025 figures from the 2025 IRS instructions and publications.

SEE Part 3 specific areas of representation

Mon, 21 Sep 2026 00:00:00 +0000

Specific Areas of Representation is worth 20 of the 85 scored questions on SEE Part 3 (about 24%). It covers clients who owe tax (collection), are being examined (audits), disagree (appeals) or want penalties or interest removed (abatement). Most questions are about a sequence: which notice, which form, which deadline.

The collection process

Stage	What happens	Your client’s options
Bill and notices	Balance due notices after assessment	Pay, request more time, set up a payment plan
Notice of Federal Tax Lien	Public claim against all the taxpayer’s property	CDP hearing on the lien filing; discharge, subordination or withdrawal requests
Notice of intent to levy	Final notice before seizure of property or wages	CDP hearing (Form 12153) within 30 days
Levy and seizure	Wages, bank accounts, property taken	Release of levy for hardship; CAP appeal

The collection statute expiration date (CSED) is generally 10 years from assessment. Some actions suspend it, including a pending offer in compromise, a pending installment agreement request and a CDP hearing.

AAIA AI auditing tools and techniques

Sun, 20 Sep 2026 00:00:00 +0000

AI Auditing Tools and Techniques is worth 21% of AAIA — around 19 of the 90 questions, the smallest domain. If you hold CISA or an equivalent designation, this is the part you already know, applied to an unfamiliar subject.

Do not over-invest here. Do not skip it either: the application of familiar method to AI is where the marks are, and it differs from ordinary IS audit in specific ways.

AAIR AI risk program management

Sun, 20 Sep 2026 00:00:00 +0000

AI Risk Program Management is worth 42% of AAIR — around 38 of the 90 questions and the largest domain. Structurally it is the risk management cycle you already know. The marks are in what changes at each step when the asset is a model.

Six areas.

Risk scenario identification and assessment

You need the AI threat catalogue and, critically, which lifecycle stage each threat targets — because the control and the evidence differ by stage:

AAISM AI technologies and controls

Sun, 20 Sep 2026 00:00:00 +0000

AI Technologies and Controls is worth 38% of AAISM — around 34 of the 90 questions and the largest single domain. ISACA describes it as optimising AI security through security technologies, techniques and controls tailored to AI systems.

This is the domain CISM and CISSP holders underestimate. The credential is called security management, so candidates expect governance and risk to dominate. They do not. The technical domain is the biggest, and it is the one furthest from existing experience.

AB-620 test and manage agents explained

Sun, 20 Sep 2026 00:00:00 +0000

Test and manage agents is worth 20–25% of AB-620 — roughly a fifth of the exam. It is the smallest domain, the least exciting, and the one candidates most often leave until the night before. That is a mistake, because the material is small, concrete and quick to learn, which makes it the best return on study time in the whole syllabus.

It has two halves: evaluating whether an agent is any good, and moving it between environments without breaking it.

Agentforce Specialist AI agents explained

Sun, 20 Sep 2026 00:00:00 +0000

AI Agents is worth 35% of the Salesforce Agentforce Specialist exam — around 21 of the 60 scored questions and comfortably the largest section. It covers how agents are built, how they behave, where they run and what they are allowed to do.

How an agent works

The foundation: an agent receives input, reasons about what is being asked, selects a topic, and executes actions to fulfil it.

The outline names the basic building blocks of agent script, and hybrid reasoning with its components and benefits. Hybrid reasoning combines the flexibility of a model with deterministic control, which is the central design tension of the whole product: you want an agent that can handle unanticipated phrasing but always calculates a discount the same way.

AI-103 computer vision solutions explained

Sun, 20 Sep 2026 00:00:00 +0000

Implement computer vision solutions is worth 10–15% of AI-103, and it is not the computer vision domain you may remember from older Azure AI exams. Classical model training barely features. Instead the domain is built around three things: generating and editing images and video from prompts, understanding visual content with multimodal models and Content Understanding, and applying responsible AI to visual material.

Image and video generation

You are expected to implement solutions that generate images from text prompts and reference media, and to do the same for video. Around that sit editing workflows:

AIF-C01 applications of foundation models

Sun, 20 Sep 2026 00:00:00 +0000

Applications of Foundation Models is worth 28% of AIF-C01’s scored content — around 14 of the 50 scored questions and the largest domain. This is where the exam moves from “what is a foundation model” to “how would you actually use one”.

The central decision: RAG or fine-tuning

The most tested distinction on the exam. Both adapt a general model to your needs, and they solve different problems.

	Retrieval-augmented generation	Fine-tuning
What it does	Retrieves your content at query time and gives it to the model	Adjusts the model’s weights with your examples
Good for	Current facts, changing content, citing sources	Style, tone, format, domain-specific behaviour
When content changes	Just update the index	Retrain
Cost	Lower to start, cost per query	Higher upfront, cheaper per query
Transparency	You can see what was retrieved	Opaque

The rule to carry into the exam: if the requirement involves information that changes, or answers that must be grounded in your documents, it is RAG. If it involves the model behaving differently — a consistent tone, a house format, domain vocabulary — it is fine-tuning.

CY0-001 AI-assisted security explained

Sun, 20 Sep 2026 00:00:00 +0000

AI-assisted security is worth 24% of CY0-001 — around fourteen questions. This is the mirror image of the securing domain: here the model is the tool you defend with, not the asset you defend. The domain also covers the other side of that coin, which is attackers using the same technology against you.

Keep the direction straight. A question about protecting a model belongs to securing AI systems; a question about a model protecting something else belongs here.

Databricks Data Analyst: dashboards and Genie

Sun, 20 Sep 2026 00:00:00 +0000

Creating Dashboards and Visualizations (16%) and Developing, Sharing, and Maintaining AI/BI Genie spaces (12%) are 28% of the Data Analyst Associate exam between them — around 12 of the 45 scored questions and the largest block on the paper.

This is also the most distinctly analyst content, and Genie is the section most candidates have never touched.

Dashboards and visualisations (16%)

Choosing the right visualisation for what you are showing:

Databricks Data Engineer: Lakeflow Jobs and CI/CD

Sun, 20 Sep 2026 00:00:00 +0000

Working with Lakeflow Jobs (16%) and Implementing CI/CD (10%) are 26% of the Databricks Data Engineer Associate exam between them — around 12 of the 45 scored questions. Both are about making pipelines run reliably without someone watching.

Lakeflow Jobs (16%)

Databricks’ orchestration layer. The exam expects operational familiarity.

Jobs and tasks. A job contains one or more tasks. Tasks can run notebooks, scripts or queries, and the value of splitting work into tasks is that each can succeed, fail and retry independently.

Databricks DE Pro: ingestion, manipulation, modelling

Sun, 20 Sep 2026 00:00:00 +0000

Data Ingestion & Acquisition (12%), Data Manipulation (12%) and Data Modelling (5%) are 29% of the Databricks Data Engineer Professional exam between them in the outline that applies from 9 October 2026: around 17 of the 60 scored questions.

The October 2026 outline reshaped all three. Ingestion grew from 7% and absorbed the old Data Sharing and Federation section. Data Transformation, Cleansing, and Quality became Data Manipulation and gained semi-structured data and AI functions. Modelling now names table layouts, materialised views and metric views. The exam format page lists every move.

Databricks GenAI assembling and deploying apps

Sun, 20 Sep 2026 00:00:00 +0000

Assembling and Deploying Apps is worth 22% of the Databricks Generative AI Engineer Associate exam — around 10 of the 45 scored questions and the second largest section. Together with application development it makes building and shipping 52% of the exam.

This section is about turning something that works in a notebook into something that runs.

Packaging an application

Assembling the components — retriever, chain, prompts, model configuration — into a deployable artefact.
Managing dependencies so the deployed version behaves like the one you tested.
Configuration that differs by environment, kept out of the code.

MLflow

Central to this section and to the platform generally:

Databricks ML Associate workflows and deployment

Sun, 20 Sep 2026 00:00:00 +0000

ML Workflows (19%) and Model Deployment (12%) are 31% of the Databricks Machine Learning Associate exam between them — around 15 of the 48 scored questions. Workflows is the larger of the two; deployment is the smallest section on the exam at about six questions.

ML workflows (19%)

Turning ad-hoc experimentation into something repeatable.

Structuring the process. The end-to-end sequence: ingest, explore, prepare, engineer features, train, evaluate, register, deploy, monitor. Know which stage a described activity belongs to — questions often hinge on identifying that a step has been skipped or done out of order.

Databricks ML Professional: model deployment

Sun, 20 Sep 2026 00:00:00 +0000

Model Deployment is worth 12% of the Databricks Machine Learning Professional exam — around 7 of the 59 scored questions and by far the smallest section.

This surprises people. “Professional” and “production” suggest serving infrastructure should dominate. It does not: ML Ops at 44% is where production actually lives on this exam, and deployment is the narrower question of how a model is made available.

Cover it properly in a few focused hours. Do not give it a week.

Is AI-901 hard? An honest look at the difficulty

Sun, 20 Sep 2026 00:00:00 +0000

AI-901 is harder than the fundamentals label suggests, and harder than the AI-900 it replaced — but only for some people. If you write Python and have clicked around Azure, most candidates find it very manageable. If you come from a business, project or sales background, this exam now asks noticeably more of you than the old one did.

The single fact that settles the question: Microsoft’s audience profile states you need “knowledge of Python coding syntax and programming techniques” and that you “should be familiar with REST APIs, SDKs, and CLIs”. The retired AI-900 explicitly said the opposite — that data science and software engineering experience were not required. That sentence is gone.

MLA-C02 deployment and orchestration

Sun, 20 Sep 2026 00:00:00 +0000

Deployment and Orchestration of ML and AI Workflows is worth 24% of MLA-C02. AWS frames it as choosing deployment infrastructure and endpoints, provisioning compute, configuring auto scaling, setting up CI/CD to automate workflows, and building agentic workflows.

This is the domain where the exam’s cost instinct is strongest. Nearly every question offers a technically workable answer that costs more than necessary.

Choosing an inference approach

The most tested decision in the domain. Match the pattern to the requirement:

N10-009 network operations explained

Sun, 20 Sep 2026 00:00:00 +0000

Network operations is worth 19% of CompTIA Network+ N10-009 — around 17 of the 90 questions. It covers running a network once it exists: watching it, documenting it, keeping it available, and changing it safely.

It is the least glamorous domain and one of the easiest to score in, because the expected answers follow consistent patterns.

Monitoring

SNMP — how devices report status, and the distinction between polling and traps, where the device reports an event rather than waiting to be asked.
Syslog — centralised logging, and severity levels.
Baselines. The idea that underpins the domain: you cannot say performance has degraded without a record of normal. Any question about detecting a gradual change wants a baseline.
NetFlow and traffic analysis for understanding what is consuming bandwidth.
Alerting on conditions that require action rather than on everything.

Log aggregation: syslog collector and SIEM

Objective 3.2 lists log aggregation as a monitoring method and names two parts: a syslog collector and a SIEM. The reason for both is the same. Individual devices keep only limited local logs, so you bring the logs of every router, switch, firewall, server and access point into one place where they can be searched, retained and correlated.

SC-500 secure compute and AI explained

Sun, 20 Sep 2026 00:00:00 +0000

Secure compute is worth 20–25% of SC-500 and contains the material that makes this exam different from the AZ-500 it succeeded. It has three parts: security for AI, security for servers and VMs, and security for application platform services. The AI part is new, it is substantial, and no AZ-500 resource covers it.

Implement security for AI

Eleven objectives. They sort into three ideas.

Agents are identities you govern. Microsoft Entra Agent ID gives an agent its own identity, which means the familiar identity toolkit applies to it:

Spark Developer: SQL, tuning and streaming

Sun, 20 Sep 2026 00:00:00 +0000

Using Spark SQL (20%), Troubleshooting and Tuning (10%), Structured Streaming (10%), Spark Connect (5%) and the Pandas API (5%) are 50% of the Spark Developer Associate exam between them — around 22 of the 45 scored questions.

Using Spark SQL (20%)

The same engine, a different surface.

Writing SQL against DataFrames registered as temporary views.
Moving between the two: spark.sql() returning a DataFrame, and a DataFrame registered for SQL.
Standard analytical SQL: joins, aggregations, CTEs, window functions.

The key insight the exam tests: SQL and the DataFrame API compile to the same plan. Neither is inherently faster. Choose for clarity — SQL is often clearer for complex joins and aggregations, the DataFrame API for programmatic construction.

200-901 deployment and security explained

Fri, 09 Oct 2026 00:00:00 +0000

Application Deployment and Security is 15% of 200-901. It covers where and how applications run, how code gets from a repository to production, and the basic security knowledge a developer is expected to have. Several objectives start with interpret or utilize, which means Dockerfiles, short Python tests and bash commands do appear.

Deployment models

Model	Meaning
Private cloud	Infrastructure used by one organisation
Public cloud	Shared provider infrastructure, pay for use
Hybrid cloud	Private and public combined
Edge	Compute close to where data is produced

Edge computing reduces latency, saves bandwidth by processing data locally and can keep working when the link to a central site is poor. A factory sensor system that must react in milliseconds is the textbook scenario.

AB-100 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

You can book the AB-100 exam without holding any other certification, but you cannot earn the Agentic AI Business Solutions Architect Expert credential from it alone. Microsoft requires you to pass AB-100 and also to earn at least one qualifying associate-level certification. The order does not matter for the exam itself, only for when the credential is awarded.

Qualifying associate certifications

Microsoft’s certification page lists these associate-level routes. Each is itself an exam you must pass:

AB-730 agents and Cowork explained

Fri, 09 Oct 2026 00:00:00 +0000

Drive business outcomes by using agents is the new AB-730 domain, worth 20–25% of the exam from October 20, 2026. It tests when a prebuilt agent helps, how to choose between a prebuilt and a custom agent, and how to delegate a multi-step task to Microsoft Copilot Cowork and then review what comes back. As a user you apply agents; you are not expected to engineer them.

The outline has two groups: select and apply prebuilt Microsoft 365 Copilot agents, and perform business tasks by using Microsoft Copilot Cowork.

AB-731 exam tips: how to gain marks

Fri, 09 Oct 2026 00:00:00 +0000

The best AB-731 exam tips are about reading: find the business need, find the constraint, and pick the answer a sensible leader would defend. The exam is 45 minutes of scenarios with plausible-sounding options, so most lost marks come from choosing an answer that is true but not the best fit. These tips cover the sitting; for preparation, start with the study plan.

Pace the 45 minutes

Microsoft does not publish the number of questions, so work in rhythm instead of arithmetic. Check the clock at the quarter marks. Answer on the first pass, flag doubtful items, and use leftover time to revisit them. Some formats, such as statement sets, may not let you go back once answered, so decide those carefully the first time.

AZ-305 infrastructure solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

Design infrastructure solutions is the largest AZ-305 domain at 30–35%. It spans four areas: compute, application architecture, migrations and networking. Together they make up about a third of the exam, so this is where preparation pays most. Every question gives requirements and asks for a recommendation, and the winning answer is the one that meets them with the least operational overhead.

Design compute solutions

Option	Choose it when
Virtual machines and scale sets	You need OS control, lift-and-shift, or software that cannot be containerised
App Service	You run a web app or API and want a managed platform with slots and autoscale
Azure Container Apps	You run containers or microservices, want scale to zero and event-driven scaling, and do not want to run Kubernetes
Azure Kubernetes Service	You need the full Kubernetes API, custom operators or fine control of the cluster
Azure Container Instances	You run short-lived or simple containers without orchestration
Azure Functions	Short, event-triggered code where you pay per execution
Azure Batch	Large numbers of independent or tightly coupled compute tasks

When a scenario says “minimal management”, move up the table toward the managed options. When it says “needs a specific kernel module” or “licensed software tied to a VM”, move down. Specify the components too: VM size family for memory or compute-heavy loads, GPU sizes for rendering and training, and Spot VMs for interruptible work.

AZ-400 security and compliance explained

Fri, 09 Oct 2026 00:00:00 +0000

Develop a security and compliance plan is 10-15% of AZ-400. It covers how pipelines authenticate, where secrets live, and how code and containers are scanned for vulnerabilities. The recurring theme is to remove long-lived secrets and to shift security checks into the pipeline. If a question offers a secretless option that fits, it is usually the right one.

Authentication and authorization

Service principals, managed identities and federation

Option	What it is	Prefer when
Service principal with a secret or certificate	An app identity in Microsoft Entra ID with a credential you manage	Nothing else is available; you accept rotation
Managed identity (system-assigned)	An identity tied to one Azure resource’s lifecycle	Code running on that resource calls other Azure services
Managed identity (user-assigned)	A standalone identity that several resources can share	Several resources need the same permissions, or the identity must outlive a resource
Workload identity federation (OpenID Connect)	A service principal or managed identity that trusts tokens from GitHub or Azure DevOps	A pipeline deploys to Azure and you want no stored secret

A managed identity works only for resources running in Azure. A hosted GitHub runner is not an Azure resource, so it cannot use one directly; federation is the answer there. Self-hosted agents on an Azure VM can use the VM’s managed identity.

CCNA 200-301 services and security explained

Fri, 09 Oct 2026 00:00:00 +0000

Network Services and Security is 20% of the v2.0 CCNA. It covers securing management access with local users and AAA, copying files securely, NAT and PAT, DNS record types, IPsec VPN concepts, IPv4 ACLs and Layer 2 protections. It is the most configuration-oriented domain, so the exam expects you to recognise correct command syntax and to predict what a configuration will allow or block.

Management access and file transfer

Topic	What to know
Local users	`username` with a secret; `login local` on the lines
AAA	The device is the AAA client; the server is TACACS+ or RADIUS
TACACS+	TCP port 49, encrypts the full payload, separates authentication, authorization and accounting
RADIUS	UDP, encrypts only the password, combines authentication and authorization
Secure file transfer	SFTP and SCP run over SSH; TFTP and FTP send data in clear text

Choose TACACS+ when the question is about per-command authorization for administrators, and RADIUS when it is about network access such as wireless or 802.1X.

Certifications a student can earn pre-graduation

Fri, 09 Oct 2026 00:00:00 +0000

A student can realistically earn a fundamentals certification before graduating: AZ-900 for cloud, SC-900 for security, AI-901 or AIF-C01 for AI, and Network+ N10-009 for networking. Associate and professional exams assume you have done the job, so they are the wrong target for most students. The right first exam depends on the field you are studying and on whether you can already read a short Python script.

DP-600 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

DP-600 has no formal prerequisite: Microsoft does not require another certification or course before you book it. The exam does assume working knowledge, though. The audience profile expects you to query and analyse data with SQL, KQL and DAX, and to have experience designing, creating and managing analytical assets such as semantic models, warehouses and lakehouses. This page turns that into a checklist you can test yourself against.

What Microsoft says about the role

The official audience profile describes a candidate with subject matter expertise in designing, creating and managing analytical assets. Your responsibilities include preparing and enriching data for analysis, securing and maintaining analytics assets, and implementing and managing semantic models. You work with stakeholders on business requirements and partner with architects, analysts, engineers and administrators. You should be able to query with SQL, KQL and DAX.

DP-700 exam tips: how to pass first time

Fri, 09 Oct 2026 00:00:00 +0000

The most useful DP-700 tip is to treat almost every question as a tool-choice problem: find the one requirement in the scenario that rules out the obvious option. The exam is built on decisions such as pipeline versus notebook, lakehouse versus warehouse, Eventstream versus Spark streaming, schedule versus event trigger. Knowing each feature in isolation is not enough; you must know what separates it from its neighbour.

Underline the deciding word

Scenarios hide the answer in a phrase. Typical deciding words:

DP-750 pipelines and workloads explained

Fri, 09 Oct 2026 00:00:00 +0000

Deploy and maintain data pipelines and workloads is worth 30–35% of DP-750, tied with data preparation as the heaviest domain. It covers designing pipelines, scheduling them with Lakeflow Jobs, moving code through environments with Git and bundles, and then keeping everything fast, cheap and observable. It is the production half of the role, and candidates who only develop in notebooks find it the hardest.

Design and implement data pipelines

You are asked to decide the order of operations (ingest, validate, transform, publish) and to choose between a notebook pipeline and Lakeflow Spark Declarative Pipelines. Declarative pipelines suit table-to-table transformations where you describe the result and let the engine work out dependencies, orchestration and incremental processing, with expectations for quality. Notebooks suit custom logic, calls to external systems or steps that do not map to tables. Many designs combine both inside one job.

DP-800 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

DP-800 has no formal prerequisite: Microsoft does not require another certification before you book it. What it does assume is experience. The audience profile expects you to write T-SQL, develop databases on Microsoft SQL platforms, be familiar with CI/CD practices in GitHub and AI-assisted development tools, and understand AI concepts such as embeddings, vectors and models. If any of those are missing, fill the gap before you start the study plan.

PL-300 manage and secure Power BI explained

Fri, 09 Oct 2026 00:00:00 +0000

Manage and secure Power BI is worth 15–20% of PL-300. It covers what happens after the report is built: creating workspaces, publishing and distributing content, keeping data fresh, and controlling who can see what. It is the smallest domain, but its questions are among the most predictable, because they test rules that have one right answer.

Create and manage workspaces and assets

Workspace roles

Role	Can do
Admin	Everything, including managing the workspace and all access
Member	Edit content, publish and update the app, and add people with equal or lower roles
Contributor	Create, edit and delete content and publish to the workspace; cannot manage access
Viewer	View and interact with content; cannot edit

A person who must edit reports but not manage access is a Contributor. A person who only consumes reports is a Viewer. Know the boundaries between neighbouring roles, as questions are built on them.

SC-200 KQL explained

Fri, 09 Oct 2026 00:00:00 +0000

For SC-200 you need to read and write short KQL queries: filter a table, pick columns, count and group, join two tables and complete a query with a missing operator. KQL is not a separate domain, but it supports advanced hunting, custom detection rules, Sentinel analytics rules, hunting queries and workbooks, so it affects well over half the exam. This guide covers the patterns that matter.

How a KQL query reads

A query starts with a table and passes rows through operators separated by a pipe. Each operator receives the output of the one before it.

SC-300 identity governance explained

Fri, 09 Oct 2026 00:00:00 +0000

Plan and automate identity governance is worth 20–25% of SC-300. It covers entitlement management (access packages), access reviews, privileged access with Privileged Identity Management (PIM) and emergency access accounts, and monitoring identity activity with logs, KQL and workbooks. The principle behind all of it is that access should be requested, time-limited and reviewed.

Entitlement management

Entitlement management bundles access and automates its lifecycle.

Object	Role
Catalog	A container of resources (groups, apps, SharePoint sites) and of delegated owners
Access package	A bundle of resource roles that a user requests as one item
Policy	Who can request, approval stages, expiry, and access reviews
Connected organization	A partner whose users may request packages
Terms of use	An agreement users must accept

When an external user’s last access package expires, the account can be removed automatically; that is the lifecycle answer for guests. Delegated catalog owners let business owners manage their own packages without a tenant admin.

SC-401 insider risk management explained

Fri, 09 Oct 2026 00:00:00 +0000

Insider Risk Management (IRM) in Microsoft Purview detects risky activity by people inside the organisation, such as data theft before resignation or leaks to personal accounts, and routes the findings to investigators. It is one of the biggest topics in the 30–35% domain Manage risks, alerts, and activities. The exam asks you to pick the right template, set up the signals it needs, and work an alert through to a case.

SC-900 Microsoft compliance solutions explained

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft compliance solutions is 20–25% of SC-900. It covers the Service Trust Portal and Microsoft’s privacy principles, compliance management in the Microsoft Purview portal (Compliance Manager and compliance score), information protection and data lifecycle management (classification, sensitivity labels, data loss prevention, retention, records management), and insider risk, eDiscovery and audit. The exam asks you to match a data-protection need to the right Purview feature.

Service Trust Portal and privacy

The Service Trust Portal publishes Microsoft’s audit reports, compliance documents and certifications, so customers can check how Microsoft services meet standards. Microsoft’s privacy principles are control, transparency, security, strong legal protections, no content-based targeting and benefits to you. Know that customers control their data and Microsoft is transparent about its use.

SY0-801 security operations explained

Fri, 09 Oct 2026 00:00:00 +0000

Security Operations is the largest SY0-801 domain at 27%. It has eight objectives (4.1 to 4.8) covering the day-to-day work of a security administrator: applying controls, managing assets and vulnerabilities, monitoring, identity and access, automation, incident response and investigation. It is also where performance-based questions are most likely to ask you to apply a concept rather than recall one.

4.1 Mitigating controls

Topics include segmentation, access controls, hardening, sandboxing, deception technology (honeypot, honeynet, honeyfile, honeytoken, canary account), MDM, application allow and block lists, IDS and IPS variants (NIDS, NIPS, HIDS, HIPS, WIPS), firewalls (WAF, UTM, layer 4 and layer 7), content filtering, DLP, EDR and XDR, network access control (802.1X, captive portals, posture checks), secrets scanning, input validation, code signing, Group Policy and SELinux.

AZ-802 DNS and DHCP explained

Tue, 29 Sep 2026 00:00:00 +0000

Implement and manage an on-premises and hybrid networking infrastructure is worth 10–15% of AZ-802, and in practice it is two services: DNS and DHCP. Name resolution covers AD integration, zones and records, forwarding, hybrid resolution with Azure, DNS policies and DNSSEC. IP addressing covers DHCP scopes, reservations, failover and troubleshooting. Remote access, VPN, IPAM and the other networking roles from AZ-800 are not in this outline.

DNS

AD integration

An AD-integrated zone is stored in Active Directory rather than a file, so every DC running DNS can update it, and it supports secure dynamic updates, where only authenticated domain members can register records. Its replication scope decides where copies live:

AB-900 exam tips: how to gain marks

Mon, 28 Sep 2026 00:00:00 +0000

The best AB-900 exam tips are about reading, not memorising: find the verb in each question, match it to the right tool, and do not lose time on any single item in a 45-minute exam. These tips cover the sitting itself. If you are still planning your preparation, start with the study plan.

Pace the 45 minutes

Microsoft does not publish the number of questions, so you cannot calculate an exact time per question in advance. Check the count on the first screen, divide, and set yourself checkpoints at a quarter, half and three quarters. Two rules keep you on schedule:

AI-300 GenAI quality and observability

Mon, 28 Sep 2026 00:00:00 +0000

Implement generative AI quality assurance and observability is worth 10–15% of AI-300. It asks two questions about a generative AI application or agent: is it good enough to ship, and can you see what it is doing once it runs? The first is answered with evaluations, the second with monitoring, tracing and cost tracking in Microsoft Foundry.

Evaluation

Test datasets and data mapping

An evaluation runs over a dataset, typically a JSONL file where each row holds a query, the response, the retrieved context and sometimes a ground-truth answer. Evaluators expect named inputs, so data mapping tells each evaluator which column is the query, which is the response and which is the context. A wrong mapping is a classic reason for evaluations that run but produce meaningless scores.

AZ-104 virtual networking explained

Mon, 28 Sep 2026 00:00:00 +0000

Implement and manage virtual networking is worth 15–20% of AZ-104, but it costs candidates more marks than its weight suggests. It covers three areas: building virtual networks and routing between them, securing access to them, and name resolution and load balancing. Questions often combine several settings in one scenario and ask why traffic does not flow.

Virtual networks and subnets

A virtual network has one or more address spaces, split into subnets. Azure reserves five addresses in every subnet: the first four and the last. A /29 therefore has eight addresses but only three usable ones. Expect at least one question that depends on that arithmetic.

Is AZ-900 hard? Exam difficulty explained

Mon, 28 Sep 2026 00:00:00 +0000

AZ-900 is one of the easier Microsoft exams, but it is not a formality. It requires no hands-on skill and no code, and most candidates with a few weeks of preparation pass. What makes it harder than people expect is breadth: dozens of service names, many of which have a close neighbour, tested in short scenarios against a 45-minute clock. People who fail usually did not underestimate the depth; they underestimated the vocabulary.

AI-200 secure, monitor and troubleshoot

Sun, 27 Sep 2026 00:00:00 +0000

Secure, monitor, and troubleshoot Azure solutions is worth 20–25% of AI-200. It has four objectives in two pairs: keeping secrets and settings out of code with Azure Key Vault and Azure App Configuration, and seeing what a distributed app is doing with OpenTelemetry tracing and KQL queries. The objective list is short, so each objective carries a lot of weight.

Secure secrets with Key Vault

Access without a credential

The exam’s favourite pattern: the app gets a managed identity, the identity gets a Key Vault role such as Key Vault Secrets User, and the code uses DefaultAzureCredential with the SecretClient from the SDK. No password, client secret or connection string is stored anywhere. When an option removes a stored credential, it is usually the right one.

Is SEE Part 2 hard? Difficulty explained

Mon, 21 Sep 2026 00:00:00 +0000

Yes — for most candidates SEE Part 2 is hard, because it covers material many preparers rarely touch: partnership and S corporation basis, corporate distributions, depreciation elections, trusts, exempt organisations and farm returns. The format is identical to the other parts, and the pass mark is the same scaled 500, but the content is further from the day-to-day work of someone who mostly prepares individual returns. It is very passable with six to eight weeks of structured work; it is not passable by relying on experience alone.

SEE Part 1 taxation explained

Mon, 21 Sep 2026 00:00:00 +0000

Taxation is worth 15 of the 85 scored questions on SEE Part 1, about 18%. It covers the taxes that sit on top of ordinary income tax, and the rules for paying them on time: alternative minimum tax, self-employment tax, household employment taxes, net investment income tax, Additional Medicare Tax, underpayment penalties and interest, and special provisions for clergy, the military and decedents. The questions are about structure: who owes the tax, on what base, and on which form.

SEE Part 3 filing process explained

Mon, 21 Sep 2026 00:00:00 +0000

Filing Process is the smallest SEE Part 3 domain, with 14 of the 85 scored questions (about 16%). It covers three things: accuracy (not trusting software blindly), record keeping and data security, and electronic filing — how a practice becomes an authorised IRS e-file provider, what an Electronic Return Originator must do, and what happens when the rules are broken. The main sources are Publication 3112 and Publication 1345.

Accuracy

Software does the arithmetic, but the preparer signs the return. The outline expects you to review results rather than rely on them: check for duplicated entries (a W-2 keyed twice, a 1099 imported and entered by hand), carryovers that did not transfer, and figures that do not make sense against the prior year. Circular 230’s due diligence standard applies to the finished return, not to the software.

AAIA prerequisites: CISA or equivalent

Sun, 20 Sep 2026 00:00:00 +0000

AAIA has an enforced eligibility gate. You cannot certify without one of a defined set of professional designations.

What qualifies

CISA — ISACA’s Certified Information Systems Auditor — qualifies on its own.

Alternatively, one of a defined group of audit and accounting designations, held in an IT audit or IT advisory role:

CIA — Certified Internal Auditor, from the Institute of Internal Auditors
US CPA — from a US state board of accountancy
ACCA (or FCCA)
Canadian CPA
Australian CPA (or FCPA)
Japanese CPA
and further international equivalents recognised by ISACA

The qualifier on the second group matters: the designation must be held by someone working in IT audit or IT advisory. A CPA in financial reporting is not the intended candidate; a CPA leading technology audits is.

AAIR prerequisites: which certifications qualify

Sun, 20 Sep 2026 00:00:00 +0000

AAIR has an eligibility gate, but it is the most accommodating of ISACA’s three AI credentials. Where AAISM requires CISM or CISSP specifically, and AAIA requires CISA or an audit designation, AAIR accepts a broad range of risk, security, audit and accounting credentials.

Many people are already eligible without realising it.

What qualifies

ISACA’s own designations:

CISA — Certified Information Systems Auditor
CISM — Certified Information Security Manager
CRISC — Certified in Risk and Information Systems Control
CGEIT — Certified in the Governance of Enterprise IT
CDPSE — Certified Data Privacy Solutions Engineer

Plus approved external designations, including:

AAISM prerequisites: CISM or CISSP required

Sun, 20 Sep 2026 00:00:00 +0000

AAISM has something most certifications do not: a genuine, enforced prerequisite.

AAISM candidates must hold an active CISM or CISSP certification.

That is ISACA’s wording, and it is a gate rather than a recommendation. No amount of AI expertise, security experience or enthusiasm substitutes for it. If you hold neither credential, AAISM is not currently available to you.

The hard requirement

Requirement	Status
Active CISM or CISSP	Mandatory
Experience in security or advisory roles	Expected
Expertise assessing, implementing and maintaining AI systems	Expected

Note the word active. A lapsed CISM does not qualify, and your CISM or CISSP must remain active for your AAISM to stand. That means maintaining two credentials with two sets of continuing education and two renewal cycles — an ongoing obligation worth understanding before you start.

AB-620 prerequisites: what you need first

Sun, 20 Sep 2026 00:00:00 +0000

There are no formal prerequisites for AB-620. You can book it today with no certifications and nothing will stop you. But Microsoft’s audience profile for this exam is unusually long and unusually specific, and ignoring it is the main reason people fail.

Most exams assume a role. This one assumes a toolkit.

What Microsoft actually assumes

You should be familiar with:

Power Fx
Microsoft Dataverse
Microsoft Power Platform environments and components
Microsoft 365 Copilot
Microsoft Foundry
Adaptive cards

You need intermediate knowledge of generative AI concepts, including:

Agentforce Specialist testing and governance

Sun, 20 Sep 2026 00:00:00 +0000

Testing, Deployment, and Maintenance (10%) and Governance and Observability (10%) are 20% of the Salesforce Agentforce Specialist exam between them — around 12 of the 60 scored questions.

Two modest sections covering what happens either side of release: proving the agent works before it ships, and knowing whether it still works afterwards.

Testing, deployment and maintenance (10%)

Testing Center

Salesforce’s tool for testing agents, and the outline asks both how to use it and how its evaluations work.

AI-103 text analysis solutions explained

Sun, 20 Sep 2026 00:00:00 +0000

Implement text analysis solutions is worth 10–15% of AI-103 and splits cleanly in two: language model text analysis, and speech. The shift from earlier Azure AI exams matters here. You are no longer expected to reach for a dedicated endpoint for every task — a fair number of these jobs are now done with generative prompting and Foundry Tools, and part of what the exam tests is knowing when a purpose-built service is still the better answer.

AI-901 exam tips: how to gain marks on the day

Sun, 20 Sep 2026 00:00:00 +0000

These are tips for the exam itself rather than for studying — how to read AI-901’s questions, which distinctions decide marks, and how to handle the sitting. If you are still planning your preparation, the study plan is the better starting point.

Read for the constraint

Almost every implementation question is a scenario with a constraint buried in it: cost matters more than nuance, it must apply to every conversation, it must work on scanned pages, the output must be structured fields. Find that clause first, then eliminate every option that violates it. Two options usually disappear immediately.

AIF-C01 guidelines for responsible AI

Sun, 20 Sep 2026 00:00:00 +0000

Guidelines for Responsible AI is worth 14% of AIF-C01’s scored content — around 7 of the 50 scored questions. It is one of the two smallest domains, and it is the cheapest to learn: the concepts are few, they recur predictably, and the expected answers follow a pattern.

The core concerns

Concern	The question it answers
Fairness and bias	Does the system treat comparable people comparably?
Transparency	Can people tell how a decision was reached, and that AI was involved?
Explainability	Can the reasoning be described in terms a person understands?
Robustness	Does it behave safely under unexpected or hostile input?
Privacy and security	Is personal data protected, including in outputs?
Governance and accountability	Who is answerable for it?
Veracity	Is the output actually true, and how would you know?

Exam framing is almost always: a scenario describes a problem, and you identify which concern it primarily raises. Unequal outcomes between groups → fairness. Nobody can explain a decision → transparency or explainability. Confident falsehood → veracity, addressed by grounding.

CY0-001 AI governance, risk, and compliance explained

Sun, 20 Sep 2026 00:00:00 +0000

AI governance, risk, and compliance is worth 19% of CY0-001 — around eleven questions. CompTIA frames it as navigating global governance frameworks to ensure ethical and compliant AI adoption. In practice this domain asks a simple question in many forms: who is accountable, what did you write down, and who checked it?

It is also the most answerable domain on the exam, because the correct answers follow a pattern once you see it.

Databricks Data Analyst: analysing and securing

Sun, 20 Sep 2026 00:00:00 +0000

Analyzing Queries (15%), Securing Data (8%) and Data Modeling with Databricks SQL (5%) are 28% of the Data Analyst Associate exam between them — around 12 of the 45 scored questions.

Analysing is the largest of the three and the most distinctly analytical content on the exam: not writing a query, but judging whether its output is right.

Analyzing queries (15%)

Is this number plausible? The habit the section tests. An analyst who publishes a figure without sanity-checking it is the risk being examined.

Databricks Data Engineer: governance and troubleshooting

Sun, 20 Sep 2026 00:00:00 +0000

Governance and Security (15%) and Troubleshooting, Monitoring, and Optimization (10%) are 25% of the Databricks Data Engineer Associate exam between them — around 12 of the 45 scored questions.

Governance and security (15%)

Almost entirely Unity Catalog.

The object hierarchy. Catalog → schema → table or volume. Permissions are granted at each level and inherit downward, which is why granting broadly at catalog level is usually the wrong answer.

Access control. Granting and revoking privileges to users and groups, and applying least privilege. Groups over individuals — a scenario about managing access for a team wants a group.

Databricks DE Professional: monitoring and deploying

Sun, 20 Sep 2026 00:00:00 +0000

Monitoring and Alerting (10%) and Debugging and Deploying (10%) are 20% of the Databricks Data Engineer Professional exam between them — around 12 of the 60 scored questions. Both kept their weight in the outline that applies from 9 October 2026, and both are about what happens after a pipeline is written.

Monitoring and alerting (10%)

What to instrument. Three layers, and questions test whether you know they are different:

Databricks GenAI governance and evaluation

Sun, 20 Sep 2026 00:00:00 +0000

Governance (8%) and Evaluation and Monitoring (12%) are the two smallest sections of the Databricks Generative AI Engineer Associate exam — 20% between them, roughly 9 of the 45 scored questions. Small, but cheap to learn and easy marks if you spend an evening on them.

Governance (8%)

Around four questions. Four ideas cover most of it.

Access control through Unity Catalog. Who can see which data, which models, and which indexed content. The recurring scenario: retrieval must respect the permissions the user already has, so people cannot surface documents through an assistant that they could not open directly.

Databricks ML Associate prerequisites

Sun, 20 Sep 2026 00:00:00 +0000

There are no formal prerequisites for the Databricks Certified Machine Learning Associate exam. No prior certification, nothing checked at registration. Databricks recommends related training but does not require it.

What it does state is 6+ months of hands-on experience performing the machine learning tasks in the exam guide — and that expectation is realistic rather than decorative.

What is actually assumed

Python code-reading. Databricks states that all ML code on the exam is in Python. You must be able to read a snippet and say what it does — libraries, function calls, arguments, what comes back. You are not writing code under time pressure, but you will be shown it.

Databricks ML Professional prerequisites

Sun, 20 Sep 2026 00:00:00 +0000

There are no formal prerequisites for the Databricks Certified Machine Learning Professional exam. No certification required, nothing checked at registration, and the ML Associate is not needed.

What Databricks does state is 1+ years of hands-on experience performing the machine learning tasks in the exam guide — double the Associate’s six months, and the most demanding experience expectation of any Databricks certification covered here.

Why the year matters

Because 44% of the exam is ML Ops, and ML Ops is the study of things that go wrong slowly.

MLA-C02 operating, monitoring and securing

Sun, 20 Sep 2026 00:00:00 +0000

Operating, Monitoring, and Securing ML and AI Solutions is worth 24% of MLA-C02. AWS frames it as maintaining observability to optimise efficiency and cost, monitoring models, agentic workflows, data and infrastructure to detect issues, and securing AI and ML systems through access controls and compliance features.

This is the domain that separates people who have run ML in production from people who have only built it.

Monitoring models

Infrastructure monitoring will not tell you a model has gone wrong. A degraded model is up, fast and answering — just worse.

N10-009 network security explained

Sun, 20 Sep 2026 00:00:00 +0000

Network security is worth 14% of CompTIA Network+ N10-009 — around 13 of the 90 questions and the smallest domain. It is deliberately foundational: depth belongs to Security+.

Cover it properly, but proportionately. Thirteen questions does not justify three weeks.

Common attacks

Recognise them by their network signature:

Attack	What you would see
DoS / DDoS	Service unavailable under overwhelming traffic
ARP spoofing / poisoning	Traffic redirected; duplicate MAC associations
DNS poisoning	Correct name resolving to the wrong address
On-path (man in the middle)	Traffic intercepted between two parties
VLAN hopping	Traffic reaching a VLAN it should not
Rogue access point / evil twin	An unauthorised wireless network users connect to
Social engineering / phishing	The human is the vector

The exam gives a symptom and asks for the attack, so learn the signature rather than the definition.

SC-500 manage and monitor security posture

Sun, 20 Sep 2026 00:00:00 +0000

Manage and monitor security posture is worth 20–25% of SC-500. It covers knowing the state of your estate (Defender for Cloud), collecting what happens in it (Microsoft Sentinel), and the newest addition, Microsoft Security Copilot.

The distinction running through the whole domain: posture is about what could go wrong, monitoring is about what did.

Defender for Cloud

Six objectives:

Defender CSPM for identifying risk across the estate
Evaluating compliance against security frameworks
Enabling and configuring workload protection plans
Connecting AWS and GCP, plus hybrid environments
Microsoft Defender Vulnerability Management settings for Azure VMs
Defender EASM — discovering unprotected assets and vulnerabilities from the outside

CSPM versus workload protection is the pairing to know. CSPM is posture — misconfigurations, risk paths, recommendations, always looking at how things are set up. Workload protection plans are runtime threat detection for a resource type — Defender for Servers, for Storage, for Databases, for Containers, for AI Services.

Spark Developer Associate prerequisites

Sun, 20 Sep 2026 00:00:00 +0000

There are no formal prerequisites for the Databricks Certified Associate Developer for Apache Spark. No certification required, nothing checked at registration. Databricks recommends related training and 6+ months of hands-on practice with the tasks in the exam guide.

What is actually assumed

Python. All code on the exam is Python. You must read PySpark fluently — method chains, arguments, what each call returns. You are not writing code under time pressure, but the reading has to be quick and precise.

200-901 infrastructure automation explained

Fri, 09 Oct 2026 00:00:00 +0000

Infrastructure and Automation is worth 20% of 200-901, tied with APIs as the heaviest domain. It asks you to explain why network teams automate, to compare the tools, and above all to interpret: a script, an Ansible playbook, a bash script, a RESTCONF reply, a YANG model, a unified diff and a sequence diagram. You are rarely asked to write from scratch; you are asked what something does.

Why automate, and how

Manual changes are slow, inconsistent and hard to audit. Automation makes changes repeatable, testable and reviewable. Infrastructure as code (IaC) describes infrastructure in text files kept in version control, so a change is a commit that can be reviewed, tested and rolled back.

AB-100 exam tips: how to answer architect questions

Fri, 09 Oct 2026 00:00:00 +0000

The best AB-100 tip is to answer as an architect rather than a builder: identify the requirement and the constraint, then choose the lightest design that satisfies both, and reject answers that add build effort, risk or cost the scenario never asked for. The exam is about judgement across Microsoft’s AI and business-application platforms, so several options usually work; one fits best.

Read the constraint first

Almost every scenario carries a limiting phrase: shortest time to value, lowest running cost, no human in the loop, data must stay in a region, users stay in Microsoft 365. Underline it before reading the options. Then eliminate any answer that breaks it, even if it is technically excellent.

AB-730 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

AB-730 has no formal prerequisites: Microsoft lists no required certification, degree or job role, and you do not need to write code. Its audience profile does, however, describe someone who already uses generative AI tools and the core Microsoft 365 apps, so some experience is assumed by the questions. This page lists that background and how to fill gaps before you start the study plan.

What Microsoft says

The audience profile describes a candidate who has experience with generative AI-powered productivity tools, including Microsoft 365 Copilot, Copilot Cowork and prebuilt agents, and who uses AI to improve daily work and make informed business decisions. You should have experience with elements such as prompt engineering, context engineering and large language models, and with Copilot tools such as notebooks and pages. You are expected to be comfortable with Outlook, Word, Teams, PowerPoint and Excel, and with common business tasks: drafting emails, building presentations, generating images and managing documents. No coding is expected.

AB-731 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

AB-731 has no formal prerequisite: Microsoft’s pages do not require another certification or exam first. What Microsoft does state is a recommended background: experience leading adoption or change management in a business context, and familiarity with Microsoft 365 services, Microsoft Foundry and general AI capabilities. You are not expected to write any code.

What the audience profile says

Microsoft describes the candidate as a business decision-maker at any level who guides transformation and innovation in a team or organisation. The expected abilities are AI fluency, strategic vision and the ability to lead AI adoption across functions. In practice, that means the following background helps most.

AI certifications for non-technical roles

Fri, 09 Oct 2026 00:00:00 +0000

The AI certifications that suit non-technical roles are AB-730 if you use Microsoft 365 Copilot in your daily work, AB-731 if you decide how your organisation adopts AI, and AIF-C01 if your company works on AWS or you want a vendor view of AI concepts. None of the three needs code. The better-known AI-901 is no longer a safe choice for non-programmers, because it now assumes Python.

AZ-305 exam tips: how to answer design questions

Fri, 09 Oct 2026 00:00:00 +0000

The best AZ-305 tip is to treat every question as a requirements check, not a knowledge check: find the constraint that rules out most options, then pick the remaining answer with the least operational overhead. Most wrong answers on this exam are correct services chosen for the wrong requirement. These tips show how to read, rank and eliminate.

Read the requirement before the options

Scenario questions bury the deciding detail. Underline mentally:

AZ-400 instrumentation strategy explained

Fri, 09 Oct 2026 00:00:00 +0000

Implement an instrumentation strategy is the smallest AZ-400 domain at 5-10%, but it closes the DevOps loop: after you deploy, how do you know it works? It covers configuring monitoring for a DevOps environment and analysing the metrics it produces. Expect a few questions, most of them about Application Insights, Azure Monitor Logs and basic Kusto Query Language (KQL).

Configure monitoring

Azure Monitor and Logs

Azure Monitor collects metrics (numeric, near real-time) and logs (records stored in a Log Analytics workspace). To integrate with DevOps tools you send telemetry to a workspace and act on it with alerts, dashboards and workbooks, and you can trigger actions in pipelines or chat tools through action groups.

CCNA 200-301 AI and operations explained

Fri, 09 Oct 2026 00:00:00 +0000

AI, and Network Operations and Management is the smallest domain of the v2.0 CCNA at 10%, and it is new. It covers the role of agentic AI in network operations, choosing a good prompt for a generative AI system, the main network management approaches, SNMP, Ansible, and interpreting syslog messages. Most objectives are descriptive rather than command-heavy, so a few hours of focused study earns real marks.

Agentic AI and generative AI

Cisco’s description of the v2.0 exam says candidates may need to evaluate output and recommendations from agentic AI and digital network assistants. The ideas to know:

DP-600 exam tips: how to answer and manage time

Fri, 09 Oct 2026 00:00:00 +0000

The most useful DP-600 exam tip is to find the one requirement in a scenario that decides the answer: “no copy of the data”, “must roll back”, “never fall back to DirectQuery”, “only this one item”. Most questions have two plausible options, and a single constraint separates them. This page collects the habits that help most in 100 minutes.

Read the constraint first

Read the last sentence of a scenario before the body. It tells you what you must produce. Then skim the body for constraints: latency, language, security level, cost, who owns the data. Underline words such as “without copying”, “least effort”, “automatically”, “minimise” and “must not”. Eliminate every option that breaks one.

DP-700 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

DP-700 has no formal prerequisite: you can book it without holding any other certification. Microsoft does describe what a candidate should be able to do: you should have subject matter expertise with data loading patterns, data architectures and orchestration processes, and be skilled at manipulating and transforming data with SQL, PySpark and Kusto Query Language (KQL). That audience profile is the real prerequisite.

What is required and what is expected

Item	Status
Another certification	Not required
SQL	Expected; you should write joins, aggregates and window functions
PySpark	Expected; DataFrames, joins, aggregations, structured streaming basics
KQL	Expected; filters, summarize, bin, joins
Fabric experience	Expected; ideally real projects with lakehouses, warehouses and pipelines

The three languages

SQL. Used in warehouses and SQL analytics endpoints, and for security statements. You need joins, CTEs, aggregation, window functions and the security syntax for row-level security and masking.

DP-750 exam tips: how to approach the questions

Fri, 09 Oct 2026 00:00:00 +0000

The best DP-750 tip is to find the one constraint in each scenario before you read the options: cost, latency, isolation, history, least privilege, or minimal code. Most wrong answers are valid Databricks features that simply do not meet that constraint. The exam gives you 120 minutes, which is enough if you stay disciplined, and it rewards hands-on experience over memorised definitions.

Read the requirement, then the code

Many items include a snippet: a MERGE, a streaming read, a GRANT, a bundle command. Read the requirement sentence first so you know what the snippet should achieve, then check what the code actually does. Typical traps are a missing checkpoint location, a grant without USE SCHEMA, or a write mode that overwrites when the scenario wants an append.

DP-800 exam tips: how to pass first time

Fri, 09 Oct 2026 00:00:00 +0000

The best DP-800 exam tip is to identify the one constraint in each scenario that decides the answer: latency, tamper-evidence, who must see plaintext, whether a transaction may wait. Most wrong answers are real features that solve a different problem. Beyond that, these tips cover the traps that the outline makes likely.

Read the code, not the feature name

Many items show T-SQL and ask what it does. Take a few seconds to trace it: which rows come back, what the window frame is, whether CROSS APPLY or OUTER APPLY drops unmatched rows, whether a trigger handles many rows. Do not guess from keywords.

PL-300 exam tips: how to pass first time

Fri, 09 Oct 2026 00:00:00 +0000

The best PL-300 exam tip is to read the constraint before the options: almost every question hinges on one phrase such as only one active relationship, without giving workspace access, or must refresh daily. The exam tests whether you can pick the right Power BI feature for a stated requirement, so the options usually all sound plausible, and one word decides which is correct. These tips are about technique; the content is in the domain guides.

SC-200 exam tips: what to focus on

Fri, 09 Oct 2026 00:00:00 +0000

The most useful SC-200 tip is to spend your revision where the marks are: roughly 80% of the paper sits in managing the environment and responding to incidents, and KQL fluency gives you points in all three domains. After that, learn the product boundaries that questions test, such as automation rule versus playbook, custom detection versus analytics rule, and Audit versus content search.

Weight your revision

Domain	Weight	Suggested share of study time
Manage a security operations environment	40–45%	About 40%
Respond to security incidents	35–40%	About 30%
Perform threat hunting	20–25%	About 20%
KQL practice across all three		Daily, 15–20 minutes

Know the pairs that get confused

Automation rule vs playbook: rules triage inside Sentinel without code; playbooks are Logic Apps workflows that call outside systems.
Custom detection rule vs analytics rule: the first lives in Defender XDR and runs on advanced hunting tables; the second lives in Sentinel and runs on workspace tables.
Automated investigation vs automatic attack disruption: one investigates alerts and remediates; the other contains an active, high-confidence attack on its own.
Audit vs content search vs Graph activity logs: actions, items, and API calls.
Live response vs investigation package: an interactive session versus a collected snapshot.
Analytics tier vs data lake: data for active detection versus cheaper long retention.

Read the constraint

Most wrong answers are plausible but break one stated condition: no custom code, least privilege, device stays isolated, minimise cost, near real time. Underline the condition before reading the options. If two options both work, the condition decides.

SC-300 exam tips: how to approach it

Fri, 09 Oct 2026 00:00:00 +0000

The most useful SC-300 tip is to learn the handful of distinctions that the questions keep testing, and to read each scenario for the one constraint that selects the answer. Most wrong answers on SC-300 are plausible features applied to the wrong situation: the right service, but the wrong scope, licence or permission type.

Distinctions that decide questions

Pair	Rule of thumb
Delegated vs application permission	A user is present: delegated. A daemon runs alone: application, with admin consent
Eligible vs active	Eligible needs activation in PIM; active is in force now
System- vs user-assigned managed identity	System ties to one resource; user-assigned is shared and reusable
Entra joined vs registered vs hybrid joined	Corporate cloud-only; personal; corporate and in on-premises AD
PHS vs PTA	PHS validates in the cloud and survives an on-premises outage; PTA validates on-premises
Connect Sync vs Cloud Sync	Full topology and device options versus lightweight cloud-managed agents
User risk vs sign-in risk	Compromised account: password change. Suspicious sign-in: MFA
Report-only vs enabled	Report-only logs the result without enforcing
Entra role vs Azure RBAC role	Directory tasks versus resource tasks
Storage vs Log Analytics vs Event Hubs	Cheap archive, query, stream

Read the scenario for the constraint

Underline words such as least privilege, without a VPN, without helpdesk involvement, minimise administrative effort, automatically, temporary, external. Each narrows the answer. “Automatically” points to dynamic groups, access package expiry, auto-apply on a review or cross-tenant synchronization. “Temporary” points to PIM, a Temporary Access Pass or an expiring package.

SC-401 alerts, audit and eDiscovery

Fri, 09 Oct 2026 00:00:00 +0000

After policies are in place you have to see what they catch. In SC-401’s domain Manage risks, alerts, and activities, the sub-area Manage information security alerts and activities covers Purview Audit, Activity explorer, DLP and insider-risk alerts, alerts in Microsoft Defender XDR and Defender for Cloud Apps, and eDiscovery searches. The exam asks which tool answers an investigative question and what licence or role is needed.

Microsoft Purview Audit

Audit records user and admin activity across Microsoft 365. There are two levels.

SC-900 prerequisites: what you need first

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 has no formal prerequisites: anyone can book it, and you need no other certification, degree or work experience. Microsoft’s audience profile does say you should be familiar with Microsoft Azure and Microsoft 365, and that you want to understand how Microsoft security, compliance and identity solutions work across both. That familiarity is not checked at booking, but the questions assume it. This page lists what it means in practice before you start the study plan.

SY0-801 security program management explained

Fri, 09 Oct 2026 00:00:00 +0000

Security Program Management and Oversight is worth 14% of SY0-801, down from 20% in SY0-701. It has six objectives (5.1 to 5.6) on governance, risk, third parties, compliance, audits and awareness. It is vocabulary-heavy and calculation-light, which makes it a good domain to master quickly.

5.1 Governance, risk and compliance artifacts

Type	Examples
Guidelines	Benchmarks, advisories, implementation guides, reference architecture
Standards	Baselines, password, physical security, encryption standards
Procedures	Standard operating procedures, runbooks
Plans	Business continuity, disaster recovery
Policies	BYOD, acceptable use, clean desk, information security, incident response, data classification, access control, data disposal, vulnerability disclosure, privacy

Policies state intent and are mandatory; standards define specific requirements; procedures give steps; guidelines are advice.

AZ-802 storage and file services explained

Tue, 29 Sep 2026 00:00:00 +0000

Manage storage and file services is worth 15–20% of AZ-802 and has more objectives than any other domain. It covers three areas: Azure Files and Azure File Sync, Windows Server file shares, and the Windows Server storage stack from disks to encryption. Questions are about choosing the right feature for a requirement and knowing its prerequisites: what a feature needs, what it cannot do, and which edition or file system it works with.

AI-300 GenAI optimization: RAG and tuning

Mon, 28 Sep 2026 00:00:00 +0000

Optimize generative AI systems and model performance is worth 10–15% of AI-300. It covers the two main ways to make a generative AI application better once it works: improve retrieval in a RAG system, and customise the model itself through fine-tuning. The exam asks which lever fits a symptom, so the skill is diagnosis first, then the right fix.

Optimising RAG

Most RAG problems are retrieval problems. If the right passage never reaches the model, no prompt will fix the answer. The outline names four areas.

AZ-104 monitoring and backup explained

Mon, 28 Sep 2026 00:00:00 +0000

Monitor and maintain Azure resources is the smallest AZ-104 domain at 10–15%, but its questions are among the most predictable on the exam. It covers two areas: monitoring with Azure Monitor and Network Watcher, and backup and recovery with Azure Backup and Azure Site Recovery. If you know which tool answers which question, most of this domain is quick marks.

Monitor resources in Azure

Metrics and logs

Azure Monitor collects two kinds of data:

AZ-900 prerequisites: what you need first

Mon, 28 Sep 2026 00:00:00 +0000

AZ-900 has no formal prerequisites: anyone can book it, and you do not need another certification, a degree or work experience first. Microsoft’s audience profile does, however, describe candidates who “have skills and experience working with an area of IT”, such as infrastructure management, database management or software development. That background is not checked, but it is assumed by the questions. This page lists what it means in practice and how to close any gaps before you start the study plan.

Is AB-900 hard? Exam difficulty explained

Mon, 28 Sep 2026 00:00:00 +0000

AB-900 is an easy-to-moderate exam: easier than any associate exam, but harder than the old MS-900 and harder than its “fundamentals” label suggests. Nothing in it goes deep, and there is no scripting or configuration from memory. The difficulty comes from breadth and similarity — a long list of Purview and Entra tools that sound alike — and from Copilot administration, which changes faster than most study material.

What makes it hard

Many tools that sound alike

The data protection domain is worth 35–40% and names more than a dozen tools. Several pairs are easy to confuse:

Is AI-200 hard? Exam difficulty explained

Sun, 27 Sep 2026 00:00:00 +0000

AI-200 is a moderately hard associate exam: harder than a fundamentals exam, easier than an expert one, and noticeably harder for people who have never deployed code to Azure. The difficulty does not come from the number of services, which is smaller than AZ-204’s. It comes from three things: the depth expected in the data tier, the code and query snippets you have to read under time pressure, and the shortage of good study material for a new exam.

How to become an enrolled agent: SEE Part 3

Mon, 21 Sep 2026 00:00:00 +0000

To become an enrolled agent you get a PTIN, pass all three parts of the Special Enrollment Examination within three years of each other, apply on Form 23 within one year of passing the last part, pay the $140 enrollment fee and pass a suitability check covering your tax compliance and criminal background. The IRS lists no degree or experience requirement to sit the exam. Once enrolled, you renew every three years and complete 72 hours of continuing education per cycle.

SEE Part 1 advising individuals explained

Mon, 21 Sep 2026 00:00:00 +0000

Advising the Individual Taxpayer is worth 11 of the 85 scored questions on SEE Part 1, about 13%. It tests whether you can use the rules from the other domains to give advice: which filing status to choose, how to relieve a spouse of a joint liability, when to realise a gain, how to fund education or retirement, what to do after a divorce, and how and when to claim a refund. The questions are scenarios about what the client should do, not what goes on a line of the return.

SEE Part 2 exam tips: timing and traps

Mon, 21 Sep 2026 00:00:00 +0000

The most useful SEE Part 2 exam tips are about time and structure rather than tax: you have 3.5 hours for 100 questions — about 2 minutes 6 seconds each — split into three sections you cannot return to once you leave them. Budget roughly 70 minutes per section, resolve every flagged question before you close a section, take the scheduled breaks because the clock stops, read every EXCEPT question twice, and work basis calculations on paper or the whiteboard in a fixed layout. The rest of this page explains each of those.

Agentforce Specialist multi-agent orchestration

Sun, 20 Sep 2026 00:00:00 +0000

Multi-Agent Orchestration is worth 5% of the Salesforce Agentforce Specialist exam — around three of the 60 scored questions and the smallest section in the outline.

Three questions does not justify a week of study. It does justify an hour, because the concepts are self-contained and the marks are straightforward.

Salesforce names two objectives.

When a multi-agent architecture is appropriate

Given a scenario, determine whether a Multi Agent architecture is appropriate for scalability and control.

AI-103 information extraction solutions explained

Sun, 20 Sep 2026 00:00:00 +0000

Implement information extraction solutions is worth 10–15% of AI-103, and despite the modest weight it is the domain that quietly decides whether the rest of your solution works. It covers two things: building the retrieval and grounding pipeline that feeds a RAG application, and pulling structured content out of documents. Every agent that answers from your own data depends on both.

Retrieval and grounding pipelines

The objectives walk through a pipeline end to end.

AI-901 prerequisites: what you need before you start

Sun, 20 Sep 2026 00:00:00 +0000

There are no formal prerequisites for AI-901. You can book it today with no certifications, no training and no experience, and nothing will stop you. But Microsoft publishes an audience profile that describes assumed knowledge, and unlike most fundamentals exams, this one has real technical assumptions in it. Ignoring them is the usual reason a first attempt fails.

What Microsoft actually assumes

The audience profile states that a candidate is “at the beginning of your career in AI solution development” and should have:

AIF-C01 security, compliance and governance

Sun, 20 Sep 2026 00:00:00 +0000

Security, Compliance, and Governance for AI Solutions is worth 14% of AIF-C01’s scored content — around 7 of the 50 scored questions. Like responsible AI, it is small, predictable and cheap to learn.

Remember the scope limit: AWS states that implementing security protocols and developing governance frameworks are out of scope for this candidate. You need to know the concepts and who is responsible for what. You do not need to configure anything.

Databricks DE Professional: security and governance

Sun, 20 Sep 2026 00:00:00 +0000

Ensuring Data Security and Compliance (8%) and Data Governance (5%) are 13% of the Databricks Data Engineer Professional exam between them in the outline that applies from 9 October 2026: around 8 of the 60 scored questions.

This page used to cover a third section, Data Sharing and Federation. The October 2026 outline folded sharing, Clean Rooms and Lakehouse Federation into ingestion, so they are now explained on the ingestion, manipulation and modelling page.

Is AAIA hard? Exam difficulty explained

Sun, 20 Sep 2026 00:00:00 +0000

AAIA is hard for a specific and predictable reason: the domain auditors expect to dominate is the smallest, and the domain they expect to skim is nearly half the exam.

Audit technique is 21%. AI Operations is 46%. Candidates who allocate study time by familiarity study the wrong things, and that single misallocation explains most failures.

The weighting is the difficulty

Domain	Weight	How it feels to an auditor
AI Governance and Risk	33%	Familiar shape, unfamiliar content
AI Operations	46%	Largely new
AI Auditing Tools and Techniques	21%	Already yours

You are not being re-tested on auditing. You are being tested on whether you understand AI systems well enough to audit them — how they are built, how they change, how they fail and what evidence exists about them.

Is AAIR hard? Exam difficulty explained

Sun, 20 Sep 2026 00:00:00 +0000

AAIR is moderately hard, and easier than most candidates fear — provided you understand what it actually tests. The common assumption is that an AI credential demands technical knowledge. It does not. Seventy-nine per cent of AAIR is governance and risk programme management. The technical-sounding lifecycle domain is 21%.

If you work in risk and hold a qualifying designation, this is closer to a specialisation exam than a new discipline.

Is AAISM hard? Exam difficulty explained

Sun, 20 Sep 2026 00:00:00 +0000

AAISM is a hard exam with an unusually well-prepared candidate pool, and those two facts interact. Everyone sitting it already holds CISM or CISSP, so the baseline is high and the questions are written accordingly. ISACA does not have to explain what a risk threshold is.

The difficulty is not the concepts. It is the combination of ISACA’s question style with subject matter that is new to almost everyone.

What makes it hard

The technologies domain is the largest. At 38%, AI Technologies and Controls is bigger than governance or risk. Candidates expect a security management credential to be mostly policy and find the biggest domain is architecture, model lifecycle, data controls and monitoring. This mismatch of expectation is the single most common preparation error.

Is AB-620 hard? Copilot Studio exam difficulty

Sun, 20 Sep 2026 00:00:00 +0000

AB-620 is harder than most people expect from a low-code exam, and the difficulty is breadth rather than depth. Nothing on it is conceptually difficult. There is simply a lot of it, spread across Copilot Studio, the Power Platform, Foundry, Azure and two protocols that barely existed a short while ago.

The people who find it hard are not the ones who struggle with any single topic. They are the ones who know Copilot Studio well and have never touched solutions, or who know the Power Platform well and have never configured MCP.

Is CY0-001 hard? CompTIA SecAI+ difficulty

Sun, 20 Sep 2026 00:00:00 +0000

CY0-001 is moderately hard, and the difficulty is not distributed the way people expect. The concepts are approachable for anyone already working in security. What catches candidates out is the clock, the newness of the subject, and one domain worth 40% of the score.

The three real difficulties

One minute per question. Sixty questions, sixty minutes, including performance-based items that take longer than multiple choice. There is no slack. Most people who fail report running out of time rather than meeting unfamiliar material, and the usual cause is spending eight minutes on an early simulation.

Is SC-500 hard? Exam difficulty explained

Sun, 20 Sep 2026 00:00:00 +0000

SC-500 is one of the harder Microsoft associate exams, and the reason is breadth rather than depth. Identity, Key Vault, governance, storage, databases, networking, servers, containers, application services, AI security, Defender for Cloud, Sentinel and Security Copilot — all in one 120-minute paper, with no domain worth less than a fifth of the score.

Nothing on it is conceptually hard for a working Azure security engineer. There is simply nowhere to hide.

Is the Databricks Data Analyst Associate hard?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Data Analyst Associate is among the more approachable certifications covered here, but it catches out a specific group: analysts with strong SQL who assume that is the exam.

Query writing is 20%. The other 80% is the platform, dashboards, Genie spaces, governance and interpretation.

What makes it manageable

Nine sections, none above 20%. Nothing to dread, and the score report is granular.

Multiple choice only, two minutes per question. No all-or-nothing formats and comfortable pacing.

Is the Databricks Data Engineer Associate hard?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Data Engineer Associate exam is one of the more approachable certifications in this field — provided you have used the platform. It is 90 minutes, 45 scored questions, multiple choice only, with no section worth more than 22%.

The difficulty is breadth, not depth. Seven sections means you meet a little of everything and cannot skip anything.

What makes it manageable

No section dominates. The largest is 22%, the smallest 6%. There is no monster domain to fear.

Is the Databricks GenAI Engineer exam hard?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Generative AI Engineer Associate exam is moderately hard, and unusually predictable. Ninety minutes, 45 scored questions, multiple choice only, six clearly weighted sections. There are no trick formats and no hidden domains.

The difficulty is entirely about whether you have built a RAG application on Databricks. If you have, most of the exam is recognition. If you have not, no amount of reading fully substitutes.

What makes it manageable

Multiple choice only. No ordering, no matching, no drag-and-drop, no all-or-nothing partial credit traps.

Is the Databricks ML Associate exam hard?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Machine Learning Associate exam is moderately hard, and its difficulty is misread more often than most. Candidates assume a machine learning certification tests machine learning. Its largest section, at 38%, is the platform.

That single fact predicts who finds it easy. People who use Databricks daily with moderate ML knowledge generally pass. People with strong ML theory who have barely touched MLflow generally struggle.

What makes it manageable

Multiple choice only. No ordering, matching or drag-and-drop.

Is the Databricks ML Professional exam hard?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Machine Learning Professional is the hardest Databricks certification covered here, and the reason is specific: 44% of it is about problems that only appear after a model has been in production for months.

You can study model development. ML Ops largely has to be lived.

The 44/44/12 problem

Section	Weight
Model Development	44%
ML Ops	44%
Model Deployment	12%

Two sections of 26 questions each and one of seven. There is no small section to lean on and no large one to specialise in. Weakness in either 44% section is close to fatal, because passing on the strength of the other would require near-perfection.

Is the Spark Developer Associate exam hard?

Sun, 20 Sep 2026 00:00:00 +0000

The Spark Developer Associate is moderately hard, and harder than its associate label suggests for one reason: 20% of it is architecture, and the architecture questions are not definitions. They ask you to reason about how Spark actually executes work.

Candidates who have written plenty of PySpark without ever opening the Spark UI tend to find this exam more demanding than expected.

What makes it manageable

Multiple choice, two minutes per question. No format traps.

MLA-C02 prerequisites: what you need first

Sun, 20 Sep 2026 00:00:00 +0000

There are no formal prerequisites for MLA-C02 — no required certification, nothing checked at booking. But AWS publishes a target candidate description that is unusually specific, and it is the best predictor of whether you will pass.

What AWS expects

At least 1 year of experience using Amazon SageMaker AI, Amazon Bedrock and other AWS services for ML engineering
At least 1 year in a related role: backend software developer, DevOps developer, data engineer or data scientist
Experience with both traditional ML and generative AI

Plus general IT knowledge:

N10-009 network troubleshooting explained

Sun, 20 Sep 2026 00:00:00 +0000

Network troubleshooting is worth 24% of CompTIA Network+ N10-009 — around 22 of the 90 questions and the largest domain. It is also where the performance-based questions concentrate.

That weighting is the most important thing to know about this exam. Network+ is not primarily a theory test; nearly a quarter of it asks what you would do when something is broken.

The methodology

CompTIA examines its troubleshooting methodology in order, and “what should you do next” questions are answered by knowing the sequence:

200-901 network fundamentals explained

Fri, 09 Oct 2026 00:00:00 +0000

Network Fundamentals is 15% of 200-901. It is a conceptual layer: you describe and interpret, you do not configure a router. It matters because every automation script runs on a network, and the exam wants to know that you can tell whether a failure is in your code or in the path to the API. If you hold the CCNA, this domain is mostly review; if you do not, the CCNA infrastructure guide is a deeper companion.

AB-730 exam tips: how to pass in 45 minutes

Fri, 09 Oct 2026 00:00:00 +0000

The best AB-730 tips are about reading, not memorising: identify what the person in the scenario wants to do, choose the Copilot feature that does exactly that, and never skip the verification step when risk is involved. With 45 minutes and no published question count, steady pacing matters more than anything else. These tips apply to the outline in force from October 20, 2026.

Pace yourself

Microsoft does not publish the number of questions, so you cannot divide the time up front. Instead, never spend more than about a minute on a first pass. Mark the question, take your best guess, and move on. Use the remaining time to return to flagged items. Unanswered questions score nothing, so always select something.

AZ-305 prerequisites: AZ-104 and what to know

Fri, 09 Oct 2026 00:00:00 +0000

To earn the Azure Solutions Architect Expert certification you must hold the Azure Administrator Associate certification, which you get by passing AZ-104, and pass AZ-305. Microsoft states this on the certification page. The exam itself has no booking prerequisite, so you can sit AZ-305 first, but the Expert credential is awarded only once both are in place.

The certification rule

Question	Answer
Is AZ-104 required?	Yes, the Azure Administrator Associate certification is the stated prerequisite for the expert certification
Can I book AZ-305 without it?	Yes; the certification is awarded when both requirements are met
Does the order matter?	Not formally, but AZ-104 knowledge makes AZ-305 easier
Does AZ-400 also count?	No; AZ-400 awards a different certification, the DevOps Engineer Expert

If you passed AZ-104 some time ago and your associate certification lapsed, check your Microsoft Learn profile before assuming it still counts. Associate and expert certifications expire annually unless renewed for free through an online assessment.

AZ-400 prerequisites and exam tips

Fri, 09 Oct 2026 00:00:00 +0000

To earn the DevOps Engineer Expert certification, Microsoft requires that you pass AZ-400 and also hold the Azure Administrator Associate or the Azure Developer Associate certification. The Azure Developer Associate certification has since been retired, so for anyone starting now the realistic prerequisite is Azure Administrator Associate, earned with AZ-104. Microsoft also expects practical experience on both GitHub and Azure DevOps.

What Microsoft says about prerequisites

Item	What the official pages state
Required exam	AZ-400
Required associate certification	Azure Administrator Associate or Azure Developer Associate (at least one)
Azure Developer Associate status	Microsoft’s page marks the certification and its renewal assessment as retired
Experience	Administering and developing in Azure, with strong skills in at least one; implementing both GitHub and Azure DevOps solutions
Role	Developer or infrastructure administrator on a cross-functional team

Check the DevOps Engineer Expert page on Microsoft Learn before booking, because certification rules change and the page is the authority. If you hold neither associate credential, AZ-104 is the straightforward way to meet the requirement, and AZ-104 vs AZ-305 shows how it fits with the architect route.

DP-600 October 2026 update: what changed

Fri, 09 Oct 2026 00:00:00 +0000

The English version of DP-600 is updated on October 19, 2026, and Microsoft’s study guide now publishes the skills measured as of that date. The outline has three domains: Maintain a data analytics solution (25–30%), Prepare data (45–50%) and Implement and manage semantic models (25–30%). Microsoft’s page says localised versions follow later, so the outline that applies to you depends on your language and booking date.

What is officially confirmed

From Microsoft Learn:

DP-750 prerequisites: what to know first

Fri, 09 Oct 2026 00:00:00 +0000

DP-750 has no formal prerequisite certification: Microsoft does not require you to pass another exam first. What it does assume is stated in the audience profile on the official study guide: you can ingest and transform data with SQL and Python, you have used software development lifecycle practices including Git, and you are familiar with Microsoft Entra, Azure Data Factory and Azure Monitor. If any of those is new, close the gap first.

Microsoft security certification path

Fri, 09 Oct 2026 00:00:00 +0000

Start with SC-900 only if you are new to security or to Microsoft’s cloud. Then pick the role exam that matches your daily work: SC-300 for identity, SC-200 for security operations, SC-401 for data protection or SC-500 for cloud and AI workload security. There is no single ladder, because these are four different jobs.

The five exams side by side

Exam	Role it tests	Duration	Take it if
SC-900	Fundamentals of security, compliance and identity	45 minutes	You are new, or you work next to security
SC-300	Identity and Access Administrator	100 minutes	You run Microsoft Entra: Conditional Access, PIM, app registrations, governance
SC-200	Security Operations Analyst	100 minutes	You work in a SOC with Defender XDR, Sentinel and KQL
SC-401	Information Security Administrator	100 minutes	You manage sensitivity labels, DLP, retention and insider risk in Purview
SC-500	Cloud and AI Security Engineer	120 minutes	You secure Azure workloads, networks, containers and AI services

Durations come from Microsoft’s exam pages. Microsoft does not publish question counts for any of them, and prices are set per country.

Retaking a certification exam after failing

Fri, 09 Oct 2026 00:00:00 +0000

After failing a certification exam, write down the domain breakdown before closing the screen, check your vendor’s retake wait, and spend that wait on the domains that cost you marks rather than rebooking at once. The waits range from none to 90 days depending on vendor and attempt number, and almost every vendor charges the full fee again, so a rushed second attempt is a second fee spent on the same gaps.

SC-200 prerequisites: what to know first

Fri, 09 Oct 2026 00:00:00 +0000

SC-200 has no mandatory prerequisite certification: Microsoft’s exam and certification pages list none. What it does state is that candidates should be familiar with Microsoft security, compliance and identity solutions, Microsoft 365, Azure cloud services, AI agents and Copilots, and Windows, Linux and mobile operating systems. This guide turns that list into a self-check.

The official familiarity list

Area	What it means in practice
Security, compliance and identity solutions	Know what Defender, Entra, Purview and Sentinel are for
Microsoft 365	Mailboxes, SharePoint, Teams, and the admin and audit model
Azure cloud services	Subscriptions, resource groups, Log Analytics workspaces, Logic Apps
AI agents and Copilots	What Security Copilot and agents can do and where they fit in an investigation
Windows, Linux and mobile operating systems	Processes, logs, services, and what suspicious activity looks like

Skills worth having before you start

Security basics. Attack stages, common techniques such as phishing, credential theft and lateral movement, and the MITRE ATT&CK vocabulary. If these are new, the SC-900 exam is a gentle first step, though it is not required.

SC-401 protecting data used by AI services

Fri, 09 Oct 2026 00:00:00 +0000

Protect data used by AI services is the newest part of SC-401 and a sub-area of Manage risks, alerts, and activities. The idea is simple: AI assistants can only surface what users can already reach, so the controls are the ones you already know, plus Data Security Posture Management (DSPM) for AI to see what is happening and to deploy policies quickly. The exam asks about prerequisites, roles, policies and monitoring, and about which existing Purview and Microsoft 365 control limits AI exposure.

SC-900 exam tips: how to pass first time

Fri, 09 Oct 2026 00:00:00 +0000

The most useful SC-900 tip is to match every scenario to one keyword and one product: the exam rewards recognising what each Microsoft product is for, not knowing how to configure it. Most lost points come from near-identical product names, outdated names from older study material, and rushing statement questions. This page collects the habits that help most in the 45 minutes.

Read for the verb

Each scenario contains a verb that points to the answer.

AZ-802 Windows Server security explained

Tue, 29 Sep 2026 00:00:00 +0000

Secure Windows Server infrastructure is worth 10–15% of AZ-802. It has two halves: hardening the operating system, and hardening Active Directory. The OS half covers exploit protection, App Control for Business, Credential Guard, SmartScreen, security baselines with OSConfig, Windows LAPS, Defender for Servers and Windows Firewall. The AD half covers password policies, Entra Password Protection, Protected Users, domain controller hardening, delegation and authentication protocols. Questions usually describe an attack or a compliance requirement and ask which control addresses it.

Is AB-900 worth it? Who should take it

Mon, 28 Sep 2026 00:00:00 +0000

AB-900 is worth it if you administer, support or roll out Microsoft 365 in an organisation that is adopting Copilot and agents. It is a short, inexpensive fundamentals exam that never expires, and it covers the questions every Copilot rollout raises: who can see what, how data is protected, and who pays for which licence. It is less worth it if you already hold associate-level Microsoft 365 or security certifications, or if your work is building agents rather than governing them.

Is AI-300 hard? Exam difficulty explained

Mon, 28 Sep 2026 00:00:00 +0000

AI-300 is a moderately hard to hard associate exam. It is harder than DP-100 was for most candidates, because it spans two different jobs: running traditional machine learning in Azure Machine Learning, and running generative AI in Microsoft Foundry. Few people do both every day. The difficulty comes less from any single topic than from the breadth, the DevOps tooling many data scientists have not used, and questions that describe a production problem rather than a feature.

Is AZ-104 worth it? Who should take it

Mon, 28 Sep 2026 00:00:00 +0000

AZ-104 is worth it if you run, support or build on Azure infrastructure, or if you want to become an Azure architect. It is Microsoft’s associate certification for Azure administrators, it covers the day-to-day work of that role, and it is the required prerequisite for the Azure Solutions Architect Expert credential. It is less worth it if your work is mainly writing application code or building AI features, where other Microsoft exams fit better.

Is AZ-900 worth it? Who should take it

Mon, 28 Sep 2026 00:00:00 +0000

AZ-900 is worth it if you are new to Azure, work alongside Azure teams in a non-technical role, or need a recognised first cloud credential for a job, a partner requirement or a career change. It is cheap in time, it never expires, and it gives you the vocabulary every other Azure exam assumes. It is less worth it if you already work in Azure daily: then an associate certification says more about you for a similar effort.

Is AI-200 worth it? Who should take it

Sun, 27 Sep 2026 00:00:00 +0000

AI-200 is worth it if you build or maintain back-end services on Azure, and especially if your team is adding AI features to existing applications. With AZ-204 retired on July 31, 2026, it is now Microsoft’s associate certification for Azure developers, so it has taken over the role AZ-204 played on job descriptions and partner requirements. It is less worth it if you work on models and prompts rather than the systems around them.

Is SEE Part 2 worth it? Who it suits

Mon, 21 Sep 2026 00:00:00 +0000

For a preparer who works, or wants to work, with small businesses, partnerships and S corporations, SEE Part 2 is worth it — both as a step to the Enrolled Agent credential and for what the study itself teaches. On its own a Part 2 pass is not a credential; it only counts toward enrollment, and only for three years. The real question is whether the EA as a whole is worth it, and for business-focused preparers it usually is: unlimited representation rights before the IRS for roughly $1,100 in fees and a few months of study, with no degree required.

Is SEE Part 3 worth it? The EA credential

Mon, 21 Sep 2026 00:00:00 +0000

For anyone who prepares tax returns for a living, the enrolled agent credential is usually worth it, and Part 3 is the part that delivers its main benefit. Enrolled agents have unlimited rights to represent taxpayers before the IRS — in audits, collection and appeals, for any tax and any taxpayer — the same practice rights as attorneys and CPAs. The cost is modest ($317 per part, $140 to enroll) against what representation work pays. It is not worth it if you have no intention of practising tax.

SEE Part 1 specialized returns explained

Mon, 21 Sep 2026 00:00:00 +0000

Specialized Returns for Individuals is worth 11 of the 85 scored questions on SEE Part 1, about 13%. It covers three areas outside the ordinary Form 1040: estate tax (Form 706), gift tax (Form 709) and international information reporting (FBAR, Form 8938 and related forms). Many candidates have never prepared any of these, which makes it the domain where focused study gains the most per hour.

The good news is that the exam tests structure and filing rules much more than complex estate calculations. Know what is included, what is excluded, who files, when, and what happens if they do not.

Is AAIA worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

AAIA is worth it if you are a practising auditor who has been asked — or will shortly be asked — to give assurance over an AI system. It is ISACA’s AI audit credential, it builds on CISA rather than competing with it, and it addresses a problem most audit functions are currently improvising: what does sufficient, appropriate evidence look like when the subject is a model?

It is not worth it if you do not hold a qualifying designation, because you cannot certify without one.

Is AAIR worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

AAIR is worth it if you work in risk and AI has entered your register — or is about to. It is ISACA’s AI risk credential, it has the broadest eligibility of their three AI certifications, and its content maps onto a question most enterprises are currently answering badly: how do you quantify, treat and monitor risk from a system whose behaviour comes from data and changes on its own?

Is AAISM worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

AAISM is worth it if you are already a certified security manager and AI has landed in your remit. It is ISACA’s AI-centric security management credential, it sits on top of CISM or CISSP rather than competing with them, and it addresses a set of questions — who owns AI risk, what does due diligence on a model provider look like, what does an AI incident response plan contain — that most organisations are currently improvising.

Is AB-620 worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

AB-620 is worth it if you build agents on Copilot Studio for a living, or want to. It is currently the only Microsoft associate certification aimed squarely at enterprise agent building on the low-code stack, and it covers material — MCP, A2A, computer use, Fabric data agents — that very few people have yet. If your organisation runs on Microsoft 365 and the Power Platform, this is the relevant credential.

Is Agentforce Specialist worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Salesforce Certified Agentforce Specialist is worth it if you work on the Salesforce Platform and your organisation is building agents. Salesforce is pushing Agentforce hard, customers are asking for it, and very few people can yet evidence that they can build one properly.

Outside the Salesforce ecosystem it signals little. This is a platform credential by design.

Who gets the most out of it

Salesforce administrators moving into AI work. The most direct fit. You already know the platform, objects and permissions. What you add is Prompt Builder, Agent Builder and Data 360 — a genuine extension rather than a new career.

Is AI-103 worth it? Who should take the exam

Sun, 20 Sep 2026 00:00:00 +0000

AI-103 is worth it if you build, or want to build, generative AI applications and agents on Azure and you need a credible way to show it. It is a developer exam with a strong operations streak: over half the score sits in planning, managing and building agentic solutions. If your work is data science, pure infrastructure, or AI on a platform other than Azure, your time and money are better spent elsewhere.

Is AI-901 worth it? Who should take the exam

Sun, 20 Sep 2026 00:00:00 +0000

AI-901 is worth it if you are early in an AI or cloud career and want a structured, credible way to learn what building on Azure AI actually involves. It is a genuine entry point: one exam, no prerequisites, no annual renewal, and a syllabus that now matches current tooling rather than the previous generation of it. It is not worth it if you already build AI systems for a living — in that case it teaches you little and signals less.

Is AIF-C01 worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

AIF-C01 is worth it if you work around AI on AWS without building it, and you want a credible, cheap way to show you understand what you are talking about. At 100 USD, 90 minutes and no prerequisites, it is one of the lowest-cost certifications in this field, and the syllabus is a genuinely good map of modern applied AI.

It is not worth much if you already build ML systems. For that, MLA-C02 is the exam that matches your work.

Is CY0-001 worth it? Who should take CompTIA SecAI+

Sun, 20 Sep 2026 00:00:00 +0000

CY0-001 is worth it if you already work in security and AI systems have entered your scope — or are about to. It is vendor-neutral, it is one of the first certifications aimed specifically at securing AI, and the subject matter is in demand well ahead of the supply of people who understand it. It is not worth it if you are new to security, or if you want a credential that hiring managers will recognise instantly, because in 2026 many still will not.

Is MLA-C02 worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

MLA-C02 is worth it if you build and run machine learning on AWS. It is the associate-level credential that matches that job, it now covers foundation models alongside traditional ML, and while it remains in beta it costs 75 USD — half the usual associate fee and less than the foundational AIF-C01.

It is not worth it if you use AI on AWS without building it. That is a different exam.

Is N10-009 worth it? Who should take Network+

Sun, 20 Sep 2026 00:00:00 +0000

CompTIA Network+ is worth it if you work in or want to work in infrastructure. It is vendor-neutral, widely recognised by name, and it teaches the one thing almost every other IT discipline depends on: how networks actually work.

Among the certifications covered on this site it is the most conventional and the most broadly recognised. It is also the least specialised, which is both its strength and its limitation.

Is SC-500 worth it? Who should take it

Sun, 20 Sep 2026 00:00:00 +0000

SC-500 is worth it if you secure Azure environments for a living. That is not a close call, because the alternative no longer exists: AZ-500 and the Azure Security Engineer Associate certification retired on August 31, 2026, renewal assessments included. If you want a current Microsoft credential for this role, SC-500 is the only route.

The more interesting question is whether to prioritise it now, and that depends on where you are.

Is the Databricks Data Analyst Associate worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Data Analyst Associate is worth it if you work in Databricks SQL and want a credential that matches what analysts actually do. At $200 for 90 minutes it is reasonably priced, and the syllabus covers the full analyst loop — query, check, visualise, share — plus the newer Genie spaces that most analysts have not yet touched.

Its value is tied to the platform. Outside a Databricks environment it signals little.

Is the Databricks Data Engineer Associate worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Data Engineer Associate is worth it if you do data engineering on Databricks and want a credential that matches the work. At $200 for 90 minutes, with seven sections covering ingestion, transformation, orchestration, governance, CI/CD and troubleshooting, it is a fair description of the job.

Its value is tied to the platform. Outside a Databricks environment it signals little.

Who gets the most out of it

Data engineers new to Databricks. The most direct fit. The syllabus is a good map of what competence on the platform means, and the flat weighting forces breadth rather than letting you specialise early.

Is the Databricks DE Professional worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Data Engineer Professional is worth it if you own production pipelines on Databricks. It is the platform’s senior data engineering credential, and its syllabus — code quality, cost and performance, monitoring, debugging, security, governance — describes the difference between a pipeline that runs and one you can rely on.

It is not worth it if you are early on the platform. The Associate exists for that, and the Professional assumes what it teaches.

Is the Databricks GenAI Engineer cert worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Generative AI Engineer Associate is worth it if you build LLM applications on Databricks, or want to. It is one of the few certifications aimed squarely at shipping RAG applications rather than at understanding AI conceptually, and its syllabus reads like a competent project plan.

It is not worth much if you do not work on Databricks. Unlike a vendor-neutral credential, almost all of its value is tied to the platform.

Is the Databricks ML Associate worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Machine Learning Associate is worth it if you do machine learning work on Databricks. It validates platform fluency — MLflow, AutoML, Unity Catalog — alongside the modelling workflow, and at $200 for a 90-minute exam it is reasonably priced for what it covers.

Its value is tightly bound to the platform. Outside a Databricks environment, it signals very little.

Who gets the most out of it

Data scientists working on Databricks. The obvious audience. You likely know the modelling; what this validates is that you use the platform’s tooling properly rather than treating the workspace as a notebook server.

Is the Databricks ML Professional worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Machine Learning Professional is worth it if you run machine learning in production on Databricks. It is the platform’s senior ML credential, and its syllabus — half rigorous development, half ML Ops — describes the difference between building a model and being responsible for one.

It is not worth attempting without production experience. 44% of it is about problems that only appear after months of operation.

Who gets the most out of it

ML engineers who own production models. The obvious audience. Monitoring, drift, retraining and version management are your week, and the credential says so.

Is the Spark Developer Associate worth it?

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Certified Associate Developer for Apache Spark is worth it if you write Spark code. It is the one certification in Databricks’ range that tests a technology rather than a platform, and that makes it the most portable credential they offer.

At $200 for 90 minutes, with half the exam on the DataFrame API and Spark SQL, it is a fair test of whether you can actually work in Spark.

Free 200-901 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the six 200-901 domains, weighted roughly as the exam is: three on software development, four on APIs, three on Cisco platforms, three on deployment and security, four on infrastructure and automation and three on network fundamentals. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Software Development and Design

Question 1. A developer wants to try a risky change without affecting the main line of work, and merge it later if it succeeds. What should they use?

Free AB-100 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three AB-100 skill areas, weighted roughly as the real exam is: six on Plan, five on Design and nine on Deploy. These are original scenario questions written against the official skills outline. Find the constraint in each scenario before you read the options.

Plan AI-powered business solutions

Question 1. A logistics firm wants an agent to help dispatchers. Before any build, the architect must judge whether the underlying shipment data can support grounding. Which set of properties should the review check?

Free AB-730 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the four AB-730 skill areas, weighted roughly as the real exam is: six on generative AI fundamentals, four on prompts and chats, five on business content and collaboration, and five on agents. These are original scenario questions, not questions from the real exam. Read the constraint in each question, such as how often, who, or which risk, before choosing an answer.

Generative AI fundamentals

Question 1. A user asks Copilot for the sales figures in a spreadsheet on a SharePoint site she cannot open. What happens?

Free AB-731 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three AB-731 skill areas, weighted roughly as the real exam is: seven on business value, eight on Microsoft AI apps and services, and five on implementation and adoption. These are original scenario questions. Read the business need and the constraint before choosing the tool or action, and give yourself about 10 minutes.

Identify the business value of generative AI solutions

Question 1. A manager asks a generative AI assistant: Write something about our new product. The output is generic. Without changing the model or the data, what change would most improve the result?

Free AZ-305 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty design scenarios across the four AZ-305 skill areas, weighted roughly as the real exam is: five on identity, governance and monitoring, four on data storage, four on business continuity and seven on infrastructure. These are original scenario questions. Find the deciding requirement in each scenario before you look at the options.

Design identity, governance, and monitoring solutions

Question 1. A web app on App Service must read database credentials from Key Vault. Security forbids secrets in app settings and wants no credential rotation by the team. What should you recommend?

Free AZ-400 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the five AZ-400 skill areas, weighted roughly as the real exam is: three on processes and communications, three on source control, ten on build and release pipelines, three on security and compliance and one on instrumentation. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Design and implement processes and communications

Question 1. A team wants to see on a dashboard how many days bugs wait between being reported and being resolved. Which Azure DevOps feature is the best fit?

Free CCNA 200-301 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the five v2.0 CCNA domains, split roughly by weight: five on infrastructure, five on switching, four on routing, four on services and security, and two on AI and operations. These are original scenario questions. Find the constraint in each before you look at the options, and do not stop at the right answer: read why the others fail.

Network Infrastructure and Connectivity

Question 1. A subnet uses the mask 255.255.255.224. How many usable host addresses does it provide?

Free DP-600 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three DP-600 skill areas, split by weight: five on maintaining a solution, ten on preparing data and five on semantic models. They are original scenario questions. Find the deciding constraint in each scenario before you look at the options, and give yourself about 35 minutes. The outline they follow applies from October 19, 2026 (see exam format).

Maintain a data analytics solution

Question 1. Sales managers query a warehouse with T-SQL. Each manager must see only the rows for their own region, and the filter must apply to every query without changing the queries. What should you implement?

Free DP-700 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three DP-700 skill areas, split as evenly as the outline weights them: seven on implement and manage, seven on ingest and transform, and six on monitor and optimize. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Implement and manage an analytics solution

Question 1. A finance analyst with no coding experience must clean a SaaS export (rename columns, change types, merge two queries) and load it into a lakehouse table each morning. Which item is the best fit?

Free DP-750 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the four DP-750 skill areas, weighted roughly as the real exam is: three on environment setup, three on security and governance, seven on preparing and processing data and seven on deploying and maintaining pipelines. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Set up and configure an Azure Databricks environment

Question 1. Interactive development clusters are often left running overnight, and the bill is growing. Developers still want to attach notebooks to their own compute during the day. What is the simplest setting change?

Free DP-800 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three DP-800 skill areas, weighted roughly as the real exam is: seven on designing and developing database solutions, seven on securing, optimizing and deploying them, and six on AI capabilities. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Design and develop database solutions

Question 1. Three tables must draw invoice numbers from one shared, gap-tolerant counter that is not tied to any single table. Which object should you create?

Free PL-300 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the four PL-300 skill areas, weighted roughly as the real exam is: six on preparing the data, six on modelling, five on visualising and analysing, and three on managing and securing Power BI. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Prepare the data

Question 1. A Power BI report uses a very large SQL database. Users must see changes in the source within seconds and the data is too large to import. Which storage mode should you use?

Free SC-200 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three SC-200 skill areas, weighted roughly as the real exam is: nine on managing the security operations environment, seven on responding to incidents and four on threat hunting. These are original SOC scenario questions. Find the constraint in each scenario before you look at the options.

Manage a security operations environment

Question 1. A Sentinel analytics rule must detect a single high-risk sign-in pattern as fast as possible. The query reads one table and uses no joins. Which rule type should you create?

Free SC-300 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the four SC-300 skill areas, weighted roughly as the real exam is: four on user identities, six on authentication and access management, five on workload identities and five on identity governance. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Implement and manage user identities

Question 1. Users in the Marketing department change roles often. Their membership of a Marketing security group must follow the department attribute in their profile with no manual work. What should you configure?

Free SC-401 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the three SC-401 skill areas: seven on information protection, seven on data loss prevention and retention, and six on risks, alerts and activities. The three domains carry equal ranges in the outline, so the split follows that. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Implement information protection

Question 1. A bank must detect records that match actual customer account numbers stored in a database table, with minimal false positives. What should you create?

Free SC-900 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the four SC-900 skill areas, weighted roughly as the real exam is: three on concepts, five on Microsoft Entra, eight on security solutions and four on compliance solutions. These are original scenario questions. Find the requirement in each scenario before you look at the options.

Security, compliance and identity concepts

Question 1. Which Zero Trust principle says that you should limit user access with just-in-time and just-enough access?

Free SY0-801 practice test: 20 questions

Fri, 09 Oct 2026 00:00:00 +0000

Twenty questions across the five SY0-801 domains, weighted roughly as the objectives are: three on general concepts, five on threats, four on architecture, five on operations and three on program management. These are original scenario questions. Find the keyword in each scenario before you read the options, and aim for 16 or more correct.

General Security Concepts

Question 1. A software vendor publishes a hash of each download beside the file so users can confirm it was not altered. Which security property does this protect?

Microsoft data and analytics certification path

Fri, 09 Oct 2026 00:00:00 +0000

If you build reports, start with PL-300. If you build data platforms, pick DP-700 for Microsoft Fabric or DP-750 for Azure Databricks. If you design analytics solutions on Fabric, take DP-600. If you write SQL for applications that use AI, take DP-800. Microsoft has no fixed order between these five.

The five exams side by side

Exam	Certification	Platform	Take it if
PL-300	Power BI Data Analyst Associate	Power BI	You model data and build reports
DP-600	Fabric Analytics Engineer Associate	Microsoft Fabric	You design semantic models and analytics solutions
DP-700	Fabric Data Engineer Associate	Microsoft Fabric	You ingest and transform data with pipelines, notebooks and KQL
DP-750	Azure Databricks Data Engineer Associate	Azure Databricks	You engineer data on Databricks inside Azure
DP-800	SQL AI Developer Associate	SQL and AI	You develop database solutions that use vectors and AI features

Choosing between them

Analyst or engineer? PL-300 vs DP-600 explains why PL-300 is the right start for report builders and DP-600 the next step if you also own the data model and the platform around it. DP-600 vs DP-700 separates the analytics engineer from the data engineer.

Taking IT exams in English as a second language

Fri, 09 Oct 2026 00:00:00 +0000

A non-native English speaker should first check the exam’s own language list, then request any extra time before booking, and only then decide whether to sit in English. Many Microsoft exams are offered in ten to thirteen languages, while several Databricks, CompTIA and Salesforce exams are not, so the answer differs by exam. This guide uses what the exam-format and renewal pages of this site state about language, plus CompTIA’s own testing policy, and says where they are silent.

AZ-802 monitoring and troubleshooting

Tue, 29 Sep 2026 00:00:00 +0000

Monitor and troubleshoot Windows Server environments is worth 15–20% of AZ-802. It covers three areas: monitoring with Windows Server tools and Azure services, troubleshooting Windows Server itself, and troubleshooting Active Directory. It is the domain that ties the others together, because a troubleshooting question can be about DNS, Hyper-V, BitLocker or Kerberos. What it tests is method: which tool shows the problem, and which fix addresses the cause rather than the symptom.

Free AB-900 practice test: 20 questions

Mon, 28 Sep 2026 00:00:00 +0000

Twenty questions across the three AB-900 skill areas, weighted roughly as the real exam is: seven on core features, eight on data protection and governance, and five on Copilot and agents. These are original scenario questions. Read what the requirement actually asks before picking a tool.

Identify the core features and objects of Microsoft 365 services

Question 1. Your company has just bought the domain contoso-sales.com and wants users to receive email at it. Where do you add and verify the domain?

Free AZ-104 practice test: 20 questions

Mon, 28 Sep 2026 00:00:00 +0000

Twenty questions across the five AZ-104 skill areas, weighted roughly as the real exam is: five on identity and governance, four on storage, five on compute, four on networking and two on monitoring and backup. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Manage Azure identities and governance

Question 1. Users should be added to the Sales-Apps group automatically when their department attribute is set to Sales, and removed when it changes. What should you configure?

Free AZ-900 practice test: 20 questions

Mon, 28 Sep 2026 00:00:00 +0000

Twenty questions across the three AZ-900 skill areas, weighted roughly as the real exam is: six on cloud concepts, eight on Azure architecture and services, and six on management and governance. These are original scenario questions, not questions from the real exam. Read the constraint in each question — cost, location, who, or what — before picking an answer.

Describe cloud concepts

Question 1. A company replaces its own email servers with Microsoft 365. Under the shared responsibility model, which task remains the company's responsibility?

Is AI-300 worth it? Who should take it

Mon, 28 Sep 2026 00:00:00 +0000

AI-300 is worth it if you deploy, automate or run machine learning and generative AI systems on Azure, or want to move from building models into operating them. With DP-100 retired on June 1, 2026, it is Microsoft’s associate certification for this work, and the one that replaces the Azure Data Scientist Associate on job descriptions and partner requirements. It is less worth it if your work is mainly research, analysis or model development with no path to production.

Free AI-200 practice test: 20 questions

Sun, 27 Sep 2026 00:00:00 +0000

Twenty questions across the four AI-200 skill areas, weighted roughly as the real exam is: five on containers, six on data services, four on connecting services and five on security and monitoring. These are original scenario questions. Read what the requirement actually asks before picking a service.

Develop containerized solutions on Azure

Question 1. A release pipeline must deploy exactly the image that passed testing, even if someone later pushes a new image with the same tag. How should the deployment reference the image?

Free SEE Part 2 practice test: 20 questions

Mon, 21 Sep 2026 00:00:00 +0000

This free SEE Part 2 practice test has 20 questions weighted like the real exam: 7 on Business Entities and Considerations, 9 on Business Tax Preparation and 4 on Specialized Returns and Taxpayers, mirroring the 30/37/18 split of scored questions. All questions use tax year 2025 rules, as the current exam does. They are original questions, not real exam items. At real exam pace, allow about 42 minutes.

Business Entities and Considerations

Question 1. A married couple jointly own and operate an unincorporated catering business. They are its only owners, both materially participate, file a joint return and have not formed a state-law entity. How can they report the business?

Free SEE Part 3 practice test: 20 questions

Mon, 21 Sep 2026 00:00:00 +0000

Twenty original questions weighted like the real SEE Part 3: six on Practices and Procedures, six on Representation before the IRS, five on Specific Areas of Representation and three on the Filing Process. That mirrors the 26/25/20/14 split of the 85 scored questions. None of these are real exam items. Give yourself about 40 minutes — roughly the pace the real exam requires.

Practices and Procedures

Question 1. Which of the following individuals has unlimited rights to represent taxpayers before the IRS?

SEE Part 1: is the Enrolled Agent worth it?

Mon, 21 Sep 2026 00:00:00 +0000

The Enrolled Agent credential is worth it if you prepare tax returns or want to, and especially if you want to represent clients before the IRS. An Enrolled Agent is a federally authorised tax practitioner with unlimited rights to represent taxpayers before the IRS in audits, collections and appeals, for any kind of taxpayer and any tax matter. Among tax specialists, only CPAs and attorneys have the same rights. It costs around a thousand dollars in exam fees and a few months of study, with no degree or experience required to sit the exams. Passing SEE Part 1 is the first step for most candidates.

Databricks Data Analyst practice test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the eight sections of the Databricks Data Analyst Associate exam, weighted roughly as the real exam is. Querying with Databricks SQL carries the most weight, so it gets the most questions here.

Executing queries using Databricks SQL and SQL Warehouses

Question 1. A dashboard query on a SQL warehouse is slow because it scans a very large table on every run, though analysts only ever look at the last 30 days. What is the most effective change?

Databricks Data Engineer Associate test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the seven sections of the Databricks Data Engineer Associate exam, weighted roughly as the real exam is. Ingestion and transformation together carry over 40% of the score.

Databricks Intelligence Platform

Question 1. Which statement about Delta Lake is correct?

A. It requires a dedicated database server
B. It cannot support concurrent readers and writers
C. It adds ACID transactions, schema enforcement and time travel over object storage
D. It stores data only in memory

Show answer

Answer: C

Databricks Data Engineer Professional test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the nine sections of the Databricks Data Engineer Professional exam, in the order and roughly the weighting of the outline that applies from 9 October 2026. This is a significant step up from the associate exam: expect production judgement rather than syntax recall.

Developing Code for Data Processing using Python and SQL

Question 1. A streaming job must guarantee that each source record affects the target exactly once, even after a driver restart. What makes this possible?

Databricks GenAI Engineer practice test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the six sections of the Databricks Generative AI Engineer Associate exam, weighted roughly as the real exam is. Application development carries the largest share, so it gets the most questions here.

Design Applications

Question 1. Stakeholders want an assistant that answers from 40,000 internal documents updated weekly. Which design should you propose?

A. Fine-tune a model weekly on the documents
B. Place all documents in the system prompt
C. Retrieval-augmented generation over an index refreshed as documents change
D. Train a model from scratch

Show answer

Answer: C

Databricks ML Associate practice test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the four sections of the Databricks Machine Learning Associate exam, weighted roughly as the real exam is. The platform section carries the largest share, so it gets the most questions here.

Databricks Machine Learning

Question 1. What is the primary purpose of experiment tracking in an ML workflow?

A. To schedule training jobs
B. To reduce storage costs
C. To record parameters, metrics and artefacts so runs are comparable and reproducible
D. To encrypt the training data

Show answer

Answer: C

Databricks ML Professional practice test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the three sections of the Databricks Machine Learning Professional exam. Model development and MLOps each carry a large share, so they dominate here. Expect production judgement rather than algorithm theory.

Model Development

Question 1. A feature is available in the warehouse at training time but only arrives three days after the event in production. What is the consequence?

A. The model will simply be slower in production
B. The model cannot be served as evaluated; the feature is unavailable at inference time
C. Nothing, as the warehouse will backfill it
D. The model will be more accurate in production

Show answer

Answer: B

Free AAIA practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the three AAIA domains, weighted roughly as the real exam is. AAIA is an audit credential: the right answer is usually the one an auditor could evidence, not the one that sounds most thorough.

AI Governance and Risk

Question 1. An auditor is asked to assess whether AI is governed adequately. What should be examined first?

A. The accuracy of the highest-profile model
B. The GPU procurement contracts
C. The AI system inventory, with owners and risk classification
D. The data science team’s training certificates

Show answer

Answer: C

Free AAIR practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the three AAIR domains, weighted roughly as the real exam is. AAIR is about managing an AI risk programme, so the right answer usually concerns process and accountability rather than technique.

AI Risk Governance and Framework Integration

Question 1. An organisation already runs an enterprise risk management framework. How should AI risk be handled?

A. Build a separate AI risk framework with its own reporting line
B. Integrate AI risk into the existing enterprise risk framework, extending taxonomy and criteria
C. Delegate it entirely to the data science team
D. Defer until regulation is final

Show answer

Answer: B

Free AAISM practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the three AAISM domains, weighted roughly as the real exam is. AAISM sits at the intersection of security management and AI, so answers favour controls a security manager can implement and evidence.

AI Governance and Program Management

Question 1. A security manager is establishing AI security governance. What is the first deliverable?

A. A procurement standard for GPUs
B. An inventory of AI systems with owner, data used and risk rating
C. A red-teaming tool selection
D. A vendor shortlist

Show answer

Answer: B

Free AB-620 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the three AB-620 skill areas, weighted roughly as the real exam is. The exam is about making design decisions for agent solutions, so read each scenario for the constraint that rules options out.

Plan and configure agent solutions

Question 1. A department wants an agent that answers policy questions from documents updated weekly by the HR team. What is the most maintainable design?

A. Hard-code the answers as topics
B. Ground the agent on the HR document library as a knowledge source
C. Ask HR to email changes to the developer each week
D. Retrain a custom model weekly

Show answer

Answer: B

Free Agentforce practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the six sections of the Agentforce Specialist exam, weighted roughly as the real exam is. AI Agents carries over a third of the score, so it gets the largest share here.

AI Agents

Question 1. A service agent must look up an order status and, if the order is late, offer a goodwill credit up to a fixed limit. What is the correct construction?

A. A single free-form instruction asking the model to be helpful
B. A topic with scoped actions for the lookup and the credit, constrained by the limit
C. A static knowledge article describing the policy
D. A report scheduled daily

Show answer

Answer: B

Free AI-103 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions spread across the five AI-103 skill areas in roughly the same proportion as the real exam. Work through them without notes, then check your answers. If you miss more than six, go back to the domain guides before booking.

Plan and manage an Azure AI solution

Question 1. You are selecting a region for a new Azure AI Foundry deployment. The application must use a specific generative model version, and legal requires that data stays inside the EU. What should drive the region choice?

Free AI-901 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the two AI-901 skill areas, weighted roughly as the real exam is. This is a fundamentals exam, so the questions test whether you can recognise the right capability for a scenario rather than write code.

Identify AI concepts and capabilities

Question 1. A retailer wants to predict next month's sales volume per store from three years of history. Which type of machine learning is this?

A. Classification
B. Regression
C. Clustering
D. Anomaly detection

Show answer

Answer: B

Free AIF-C01 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions spread across the five AIF-C01 domains in roughly the proportions of the real exam. This is a foundational exam: it tests whether you can match a business scenario to the right AI concept or service category.

Fundamentals of AI and ML

Question 1. A logistics company wants to estimate delivery time in minutes from historical trip data. Which type of problem is this?

A. Regression
B. Binary classification
C. Clustering
D. Dimensionality reduction

Show answer

Answer: A

Free CY0-001 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the four CY0-001 domains, weighted roughly as the real exam is. Securing AI systems carries the most weight, so it gets the most questions here.

Question 1. Which characteristic of large language models most directly creates a new attack surface compared with traditional software?

A. They are written in Python
B. They require GPUs to run
C. They do not cleanly separate instructions from data in their input
D. They produce output as text

Show answer

Answer: C

Free MLA-C02 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the four MLA-C02 domains, which carry almost equal weight on the real exam. These are engineering questions: the right answer is usually the one that survives production, not the one that is quickest to build.

Data Preparation for ML and AI

Question 1. A training dataset has a feature that is missing for 60% of rows and correlates weakly with the target. What is usually the best action?

Free N10-009 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the five N10-009 domains, weighted roughly as the real exam is. Troubleshooting carries the most weight, so it gets the most questions here too.

Networking concepts

Question 1. At which OSI layer does a device that forwards frames using MAC addresses operate?

A. Layer 1
B. Layer 2
C. Layer 3
D. Layer 4

Show answer

Answer: B

Switching on MAC addresses is a layer 2 function. Layer 3 forwards on IP addresses, and layer 1 concerns physical signalling.

Free SC-500 practice test: 20 questions

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the four SC-500 skill areas, weighted roughly as the real exam is. These are scenario questions: read what the requirement actually demands before picking the control.

Manage identity, access, and governance

Question 1. A contractor needs elevated access to one subscription for four hours a month to run a maintenance task. What should you implement?

A. A permanent owner role assignment
B. A shared administrator account
C. Time-bound, approval-gated privileged access activation
D. A long-lived service principal secret

Show answer

Answer: C

Spark Developer Associate practice test

Sun, 20 Sep 2026 00:00:00 +0000

Twenty questions across the seven sections of the Associate Developer for Apache Spark exam, weighted roughly as the real exam is. DataFrame API work carries the largest share, so it gets the most questions here.

Developing Apache Spark DataFrame/DataSet API Applications

Question 1. Which of these operations triggers execution rather than only building the plan?

A. select()
B. filter()
C. count()
D. withColumn()

Show answer

Answer: C

count() is an action and forces execution. select, filter and withColumn are transformations that only extend the logical plan.

200-901 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The most useful 200-901 resources are free: Cisco’s exam topics document, the DevNet Sandbox environments, the documentation for each Cisco API and a Linux machine on which you write Python, run Git and build a container. Add realistic practice questions to find your weak domain. Be careful with older material, because the exam was published as DevNet Associate (DEVASC) and some products in older courses have been renamed or retired.

AB-100 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best AB-100 resources are Microsoft’s study guide, the product documentation named in it, and one realistic design exercise you work through yourself. Because the exam is new and product names changed in the October 14, 2026 outline, be wary of any material that predates the update. Use practice questions to find your weak area, not as a substitute for the outline.

Start with the official material

The study guide and skills outline. Three domains: Plan (25–30%), Design (25–30%) and Deploy (40–45%). Read the section “Skills measured as of October 14, 2026” and its change log. Keep it open while you study.

AB-730 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best AB-730 resources are free: Microsoft’s study guide, the free practice assessment on the certification page, the exam sandbox, the official Microsoft 365 Copilot documentation, and hands-on time with Copilot itself. Add realistic practice questions to find your weak domain. Be careful with age, because Copilot features and names change quickly and the exam outline changes on October 20, 2026.

Start with the official material

The study guide. Microsoft’s AB-730 study guide lists every skill area with its weight and the bullets under each objective, plus a change log between the old and new outline. Read it first and keep it open. It is the only authoritative list of what can be tested. Find it linked from the certification page at Microsoft Certified: AI Business Professional on Microsoft Learn.

AB-731 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best AB-731 resources are free: Microsoft’s study guide with the skills outline, the free practice assessment, the exam sandbox and the documentation the study guide links. Add realistic practice questions to find weak areas. Because the exam was refreshed in July 2026, avoid older AI-leader material that predates the Foundry Tools naming.

Official material

Resource	What it gives you
Exam AB-731 study guide	The skills measured as of July 22, 2026, with a change log
AI Transformation Leader certification page	Exam facts, languages, practice assessment, sandbox, prep video
Free practice assessment	Linked from the certification page; shows Microsoft’s question style
Exam sandbox	The exam interface with no real content
Exam AB-731 prep videos	Microsoft’s tips and strategies video
Exam Readiness Zone	Microsoft’s exam preparation shows; check for an AB-731 episode

Documentation the study guide points to

Microsoft 365 Copilot service description: what Copilot includes, and how versions differ.
Microsoft Copilot Studio documentation: agents, knowledge sources, publishing.
Azure AI Foundry documentation: models, Foundry Tools and the platform.

Skim these for concepts and vocabulary. You do not need to learn the click paths, as the exam does not test configuration.

AZ-305 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best AZ-305 resources are free: Microsoft’s skills outline, the official practice assessment, the Azure Architecture Center, the Well-Architected Framework and the Azure documentation for each service in the outline. Add realistic practice questions to find your weak domain. The thing to be careful with is outdated material: the English exam was updated on April 17, 2026, and older courses may teach previous service names and tiers.

Start with the official material

The skills outline (study guide). Microsoft publishes a study guide for AZ-305 on Microsoft Learn with the four domains and every “recommend” objective under them. The current version is dated April 17, 2026. It also carries a change log comparing the previous and current outlines. Read it before anything else.

AZ-400 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best AZ-400 resources are free: Microsoft’s study guide, the official practice assessment, the documentation of Azure DevOps, GitHub Actions and Azure Monitor, and your own pipelines. Add realistic practice questions to find your weak domain. Be careful with outdated material: the outline was updated on July 27, 2026, and older courses may use classic release pipelines and names that have since changed.

Start with the official material

The study guide. Microsoft’s AZ-400 study guide lists every skill measured as of July 27, 2026, with sub-bullets, a change log and links to documentation. Read it before anything else and keep it open.

CCNA 200-301 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best CCNA resources are Cisco’s exam topics document, a network simulator you configure every day, and practice questions that point to your weak domain. Cisco’s own exam page lists guided learning paths on Cisco U, instructor-led training, practice exams, a free trial of Cisco Modeling Labs and community support through the Cisco Learning Network. Treat the topics document as the checklist, because the exam changed substantially between v1.1 and v2.0.

Cisco CCNA certification path

Fri, 09 Oct 2026 00:00:00 +0000

Take CCNA (200-301) first. It is Cisco’s foundation exam for networking and has no prerequisites. Add CCNA Automation (200-901) afterwards if your work involves Python, APIs and network automation, or take it first if you come from software and need only enough networking to automate it.

The two exams

Exam	Certification	Duration	Take it if
200-301	CCNA	120 minutes	You configure and troubleshoot networks
200-901	CCNA Automation (formerly DevNet Associate)	120 minutes	You automate networks with code and APIs

Cisco lists no prerequisites for either. A passed CCNA is valid for three years.

DP-600 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best DP-600 resources are free: Microsoft’s study guide with the skills outline, the official practice assessment, the Fabric documentation, and a workspace in which you build things. Add realistic practice questions to find your weak domain. The thing to watch is outdated material: Fabric changes often and the English exam outline changes on October 19, 2026, so check older courses against the current objectives.

Start with the official material

Resource	What it gives you	Link
DP-600 exam page	Duration, languages, accommodations, how to schedule	Exam page
Study guide	The skills measured with weights and the change log	Study guide
Practice assessment	Free questions in the style of the exam	Practice assessment
Exam sandbox	The interface and question types	Exam sandbox
Certification page	Overview and renewal	Fabric Analytics Engineer Associate

The exam page also lists self-paced learning paths and an instructor-led course in its preparation section; browse it for the current titles rather than relying on a list that may age.

DP-700 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best DP-700 resources are free: Microsoft’s study guide with the skills outline, the official practice assessment, the Microsoft Fabric documentation, and a Fabric workspace you build things in. Add realistic practice questions to locate your weak domain. Be careful with outdated material, because Fabric changes quickly and the outline is revised, most recently with an update on October 19, 2026.

Start with the official material

The study guide and skills outline. Microsoft’s DP-700 study guide lists the three domains, each at 30–35%, with every objective and a change log. Read it before anything else and keep it open while you study.
The certification page. The Fabric Data Engineer Associate page shows the duration, languages, the practice assessment and the renewal information.
The practice assessment. DP-700 has a free official practice assessment on the certification page. Take it once at the start and once near the end.
The exam sandbox. The exam sandbox shows the interface and question types without testing content.
Exam Readiness Zone. Microsoft’s Exam Readiness Zone hosts exam prep videos for several exams; check whether a DP-700 episode is currently listed.
Microsoft Learn training. The study guide points to self-paced learning paths and instructor-led courses for the exam. Use them for structure, then practise.

Documentation worth reading

The Microsoft Fabric documentation is the primary reference. Start with What is data engineering in Microsoft Fabric?, then read the pages for each item in the outline: lakehouse and table maintenance, data pipelines, Dataflow Gen2, notebooks and Spark, Eventstream, Eventhouse and KQL, warehouse security, OneLake security, deployment pipelines and Git integration, and the Monitor hub.

DP-750 study resources: free official material

Fri, 09 Oct 2026 00:00:00 +0000

The best DP-750 resources are mostly free: Microsoft’s study guide with the skills outline, the official practice assessment, the Azure Databricks documentation, and a workspace where you build things. Add realistic practice questions to find your weak domain. The thing to be careful about is old material, because the exam uses current product names and Databricks renames and replaces features regularly.

Start with the official material

The study guide. The official DP-750 study guide lists the four domains, every objective and the change log. The version dated October 19, 2026 is the one to follow. Keep it open while you study.

DP-800 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best DP-800 resources are Microsoft’s study guide, the documentation for each SQL feature it names, and a database you build things in. Add realistic practice questions to find your weak domain. Because the exam covers features that are new and still changing, be careful with older tutorials.

Start with the official material

The study guide. It lists the three domains with weights, every sub-area, the audience profile and a change log. The current version is the one dated for October 19, 2026. Read it before anything else and keep it open while you study.

PL-300 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best PL-300 resources are free: Microsoft’s study guide with the skills outline, the official practice assessment, the Power BI documentation, and Power BI Desktop itself, which you can install at no cost. Add realistic practice questions to find your weak domain. The thing to be careful with is outdated material, because Power BI ships changes monthly and the outline has been revised, so older courses can miss features the current exam names.

SC-200 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best SC-200 resources are free: Microsoft’s study guide, the official practice assessment, the product documentation for Defender XDR and Sentinel, and a lab tenant where you run queries and build rules. Add realistic practice questions to find your weak domain. Be careful with outdated material: the English outline changes on October 21, 2026, and the current one names data lake tiers, SOC optimization, Sentinel Graph, KQL jobs, summary rules and notebooks, which older courses lack.

SC-300 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best SC-300 resources are free: Microsoft’s study guide with the skills outline, the official practice assessment and sandbox, the Entra documentation, and a test tenant where you build the things the outline names. Add realistic practice questions to find your weak domain. Be careful with old material: Azure Active Directory is now Microsoft Entra ID, and courses that use the old name throughout predate parts of the current outline.

SC-401 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best SC-401 resources are free: Microsoft’s skills outline, the official practice assessment, the Microsoft Purview documentation, and a Microsoft 365 test tenant where you configure things yourself. Add realistic practice questions to find your weak domain. The one thing to be careful with is outdated material, because the exam is revised and the October 2026 outline adds weight to AI-related objectives that older courses never covered.

Start with the official material

The study guide and skills outline. Three domains of 30–35% each. The study guide at aka.ms/SC401-StudyGuide lists every objective, the exam date it applies from and the change log. Read it before anything else and keep it open.

SC-900 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The best SC-900 resources are free: Microsoft’s study guide with the skills outline, the official practice assessment and sandbox, and the product documentation for each area. Add realistic practice questions to find your weak area. Be careful with old material, because Microsoft renamed many of these products (for example, Azure Active Directory became Microsoft Entra ID, and the compliance portal moved into Microsoft Purview), and older courses use the previous names.

SY0-801 study resources

Fri, 09 Oct 2026 00:00:00 +0000

The most important SY0-801 resource is free: the official CompTIA exam objectives document, which lists every topic, the weights and an acronym list. Build everything else around it, add hands-on lab time, and use realistic practice questions to find weak domains. One caution up front: SY0-801 had not launched when this was written, so most books and courses are still SY0-701 material.

Start with the official material

The exam objectives. CompTIA publishes the SY0-801 objectives (version 2.0) as a PDF on its Security+ pages. It contains the five domains and percentages, the numbered objectives with example topics, an acronym list, and a sample hardware and software list for a lab. Read it end to end before you buy anything, and tick off topics as you learn them.

Which security certification to take first

Fri, 09 Oct 2026 00:00:00 +0000

Take SC-900 first if you are new to security and your work or target employer uses Microsoft cloud services, take Security+ SY0-801 first if you want a vendor-neutral baseline and already work in IT, and go straight to SC-200 only if you already work with Microsoft 365 security or Azure. Nothing requires a particular order: no exam on this site lists another certification as a prerequisite for these three.

Is AZ-802 hard? Exam difficulty explained

Tue, 29 Sep 2026 00:00:00 +0000

AZ-802 is a moderately hard associate exam, and noticeably harder for people who know Windows Server only from the GUI or only from Azure. No single topic is exotic. The difficulty comes from breadth — seven domains in 120 minutes, from FSMO roles to Azure Arc to Storage Replica — from troubleshooting scenarios that cross several services, and from the fact that the exam is new, with little official practice material yet.

AB-900 study resources

Mon, 28 Sep 2026 00:00:00 +0000

The best AB-900 resources are free: Microsoft’s skills outline, the Microsoft 365, Entra, Purview and Copilot documentation on Microsoft Learn, the free practice assessment, and a trial tenant to click through. Add realistic practice questions to find your weakest domain. Be careful with older material: much of what is written about Microsoft 365 fundamentals was made for the retired MS-900 and misses Copilot, oversharing and DSPM for AI.

Start with the official material

The skills outline. Three domains, with data protection and governance the largest at 35–40%. It fits on two pages and names every tool the exam covers. Read it first and keep it open while you study. Make sure you are reading the version for October 14, 2026 onward, which Microsoft’s study guide already shows.

AZ-104 study resources

Mon, 28 Sep 2026 00:00:00 +0000

The best AZ-104 resources are free: Microsoft’s skills outline, the official practice assessment, the Azure documentation for each service in the outline, and an Azure subscription you build things in. Add realistic practice questions to find your weak domain. The one thing to be careful with is outdated material, because AZ-104 has been revised several times and older courses teach settings and service names that have since changed.

Start with the official material

The skills outline. Five domains, from 10–15% to 20–25%, with identity and compute the largest. It names every service and task the exam covers. Read it before anything else, and keep it open while you study. The current version is dated April 17, 2026.

AZ-900 study resources

Mon, 28 Sep 2026 00:00:00 +0000

The best AZ-900 resources are free: Microsoft’s skills outline and study guide, the Microsoft Learn training for the exam, the free official practice assessment, and a free Azure account to look around in. Add realistic practice questions to find your weakest domain. The main thing to be careful with is age: Azure renames services often, and older material uses names the current exam no longer does.

Start with the official material

The study guide and skills outline. Three domains — cloud concepts (25–30%), architecture and services (35–40%), management and governance (30–35%) — with every sub-objective listed. It was last revised on July 20, 2026. Read it before anything else and keep it open while you study; it is the only authoritative list of what can be tested.

Free AI-300 practice test: 20 questions

Mon, 28 Sep 2026 00:00:00 +0000

Twenty questions across the five AI-300 skill areas, weighted roughly as the real exam is: four on MLOps infrastructure, six on the model lifecycle, four on GenAIOps infrastructure, three on quality and observability and three on optimisation. These are original scenario questions. Read what the requirement actually asks before picking an answer.

Design and implement an MLOps infrastructure

Question 1. Training jobs need a machine that scales to zero nodes when no jobs are queued and up to eight nodes under load. Which compute target should you create?

AI-200 study resources

Sun, 27 Sep 2026 00:00:00 +0000

The best AI-200 resources are free: Microsoft’s skills outline, the Azure documentation for each named service, and an Azure subscription you build things in. Add realistic practice questions to find your weak domain. The one thing to be careful with is older material, because much of what is labelled AI-200 online was written for AZ-204 and only partly updated.

Start with the official material

The skills outline. Four domains, between 20% and 30% each, with the data domain the largest. It fits on two pages and names every service and technique the exam covers. Read it before anything else and keep it open while you study.

Free SEE Part 1 practice test: 20 questions

Mon, 21 Sep 2026 00:00:00 +0000

Twenty questions across the six SEE Part 1 domains, weighted roughly as the real exam’s 85 scored questions: 3 on preliminary work, 4 on income and assets, 4 on deductions and credits, 3 on taxation, 3 on advising and 3 on specialized returns. They are original questions, not real exam items, and they assume tax year 2025. Give yourself about 40 minutes (the exam allows roughly two minutes per question), closed book, calculator only.

SEE Part 2 study resources

Mon, 21 Sep 2026 00:00:00 +0000

The best SEE Part 2 study material is free and official: the PSI Candidate Information Bulletin for the syllabus, the IRS’s own sample questions, and the 2025 editions of the IRS publications and form instructions the exam is based on. The exam tests law as amended through December 31, 2025, so use the 2025 editions — older ones miss the 2025 depreciation changes. We checked every publication below on irs.gov in September 2026.

SEE Part 3 study resources

Mon, 21 Sep 2026 00:00:00 +0000

Almost everything SEE Part 3 tests is published free by the IRS. Unlike Parts 1 and 2, where commercial review courses condense a large and shifting body of tax law, Part 3 rests on a handful of stable primary documents. Read these and you have read the syllabus’s sources.

Start with the rules of the exam

The PSI Candidate Information Bulletin. It contains the Part 3 content outline, the question counts per domain (26, 25, 20, 14) and the testing rules. Every topic in the outline is fair game; treat it as a checklist.

AAIA study resources

Sun, 20 Sep 2026 00:00:00 +0000

AAIA tests audit discipline applied to AI, not AI knowledge with an audit flavour. If you already audit IT, the new material is the AI-specific subject matter and it is a four-to-six week exam. If you do not, no resource list will close that gap — the reasoning the questions reward comes from having sat on the other side of a control walkthrough.

Start with the official material

ISACA’s own study material for the certificate. As a newer credential, the official resources are the only ones written directly to the syllabus, and third-party coverage is thin.

AAIR study resources

Sun, 20 Sep 2026 00:00:00 +0000

AAIR is about running an AI risk programme. The exam rewards answers concerning accountability, escalation and what actually changes as a result of an assessment — not technical depth. Experienced risk professionals find the AI content the only new part; people without risk experience find the whole framing unfamiliar, and that is the harder gap.

Start with the official material

ISACA’s study material for the certificate. It is a newer credential with limited third-party coverage, so the official resources carry more weight than usual.

AAISM study resources

Sun, 20 Sep 2026 00:00:00 +0000

AAISM sits between security management and AI. Its largest domain is AI Technologies and Controls at 38%, so it demands more technical grounding than the other two ISACA AI certificates — but the reasoning it rewards is still managerial: which control removes an attack’s mechanism, and how you would evidence that it operates.

Start with the official material

ISACA’s study material for the certificate, which is the only resource written directly to this syllabus.

AB-620 study resources

Sun, 20 Sep 2026 00:00:00 +0000

AB-620 is a design exam. It asks which construction satisfies a stated constraint, so the fastest preparation is to build one agent end to end — with a knowledge source, an action that writes data, and an escalation to a human. Microsoft Learn covers the syllabus; the build is what makes the questions obvious.

Start with the official material

The skills outline. Three areas: planning and configuring agent solutions (30–35%), integrating and extending agents in Copilot Studio (40–45%), and testing and managing agents (20–25%). The integration area is the largest single block in the exam.

Agentforce Specialist study resources

Sun, 20 Sep 2026 00:00:00 +0000

Trailhead is free, official and sufficient for the agent content. The section that catches people is Data 360 at 20% — candidates from an AI background consistently lose marks there rather than on prompt engineering. Budget study time by the published weights, not by what feels interesting.

Start with the official material

The official exam guide. Six sections: AI Agents 35%, Prompt Engineering 20%, Data 360 Fundamentals 20%, Testing Deployment and Maintenance 10%, Governance and Observability 10%, Multi-Agent Orchestration 5%.

AI-103 study resources

Sun, 20 Sep 2026 00:00:00 +0000

The official Microsoft Learn learning path for AI-103 is the only resource guaranteed to track the current skills outline, and it is free. Everything else is supplementary. The single highest-value activity, though, is not reading at all — it is building one retrieval-augmented generation application end to end.

Start with the official material

The skills outline on the exam page. Read it before anything else, and again the week you book. It lists the exact skills measured and their weights, and it is the document the exam is written from. Course material lags it by months.

AI-901 study resources

Sun, 20 Sep 2026 00:00:00 +0000

AI-901 is the exam where free official material genuinely suffices. The Microsoft Learn path for it covers the syllabus, and paying for a course is hard to justify. What people underestimate is how much a couple of hours clicking through Azure AI Foundry improves recall of material they have only read.

Start with the official material

The skills outline. Two skill areas only: identifying AI concepts and capabilities, and implementing AI solutions with Microsoft Foundry. Well over half the weight sits in the second, which tells you where to spend your time.

AIF-C01 study resources

Sun, 20 Sep 2026 00:00:00 +0000

AWS Skill Builder’s free AI Practitioner material covers this exam adequately, and the official exam guide plus sample questions tell you the style. AIF-C01 is conceptual rather than hands-on, so extensive lab work is not required — but an hour actually using a foundation model repays itself in recall.

Start with the official material

The exam guide. Five domains: AI and ML fundamentals (20%), generative AI fundamentals (24%), applications of foundation models (28%), responsible AI (14%), and security, compliance and governance (14%). Note that the last two together are 28% — as much as the largest single domain. People routinely under-study them.

CY0-001 study resources

Sun, 20 Sep 2026 00:00:00 +0000

SecAI+ is new, so third-party material is thin and some of it is poor. The compensating advantage is that the underlying subject is well documented in free, high-quality public frameworks. CompTIA’s objectives plus the published AI security literature covers this exam better than most paid courses currently do.

Start with the official material

The exam objectives. CompTIA publishes a detailed objectives document. Since the exam is new, this matters more than usual — it is the only authoritative scope, and course authors are still catching up.

Data Engineer Professional resources

Sun, 20 Sep 2026 00:00:00 +0000

This exam assumes you have operated production pipelines. No resource list substitutes for that. If you have been on call for a job that degraded, debugged a skewed stage at an awkward hour and argued about cost attribution, the material is a refresher. If you have not, the honest advice is to get six to twelve months of that experience first.

Start with the official material

The exam guide and its weights for the outline that applies from 9 October 2026: developing code for data processing 23%, cost and performance optimisation 15%, ingestion and data manipulation at 12% each, monitoring and alerting and debugging and deploying at 10% each, security and compliance 8%, and governance and data modelling at 5% each. The October 2026 guide covers both the old and new outline, so make sure you read the new one.

Databricks Data Analyst study resources

Sun, 20 Sep 2026 00:00:00 +0000

Databricks Academy’s official learning path is free for the self-paced material and is written to the exam. The complication specific to Databricks is that no passing score is published, so a weighted practice test is the only signal you get about readiness.

Start with the official material

The exam guide on the Databricks certification page, which lists the sections and their weights. Executing queries with Databricks SQL carries 20%, dashboards and visualisations 16%, analysing queries 15%, Genie spaces 12%, platform understanding 11%, and managing, securing and importing data 21% between them.

Databricks Data Engineer study resources

Sun, 20 Sep 2026 00:00:00 +0000

Databricks Academy’s self-paced path covers the syllabus free of charge, but this exam rewards having built a pipeline that failed and been fixed. Ingestion and transformation carry over 40% between them, and both are subjects where reading conveys the vocabulary and not the judgement.

Start with the official material

The exam guide and its section weights: data transformation and modelling 22%, ingestion and loading 21%, Lakeflow jobs 16%, governance and security 15%, CI/CD 10%, troubleshooting and optimisation 10%, platform 6%.

Databricks GenAI Engineer resources

Sun, 20 Sep 2026 00:00:00 +0000

Application development carries 30% of this exam and assembling and deploying another 22%. Over half the marks therefore reward having built and shipped a retrieval application rather than read about one. Databricks Academy covers the syllabus free; the build is what makes it stick.

Start with the official material

The exam guide and its weights: application development 30%, assembling and deploying apps 22%, design applications 14%, data preparation 14%, evaluation and monitoring 12%, governance 8%.

Databricks ML Associate resources

Sun, 20 Sep 2026 00:00:00 +0000

The Databricks Machine Learning section carries 38% of this exam — more than model development at 31%. That balance surprises people who prepare as though it were a general machine learning exam. It is a platform exam: tracking, registries, feature stores and reproducibility matter as much as modelling.

Start with the official material

The exam guide and its weights: Databricks Machine Learning 38%, model development 31%, ML workflows 19%, model deployment 12%.

Databricks ML Professional resources

Sun, 20 Sep 2026 00:00:00 +0000

Model development and MLOps carry 44% each on this exam, with deployment at 12%. Nearly half of it therefore tests operating models in production — drift, promotion gates, reproducibility, rollback, feature consistency. That is experience, not reading, and it is why this exam has a reputation for catching well-prepared candidates.

Start with the official material

The exam guide and its weights, which are unusually simple: model development 44%, ML Ops 44%, model deployment 12%.

MLA-C02 study resources

Sun, 20 Sep 2026 00:00:00 +0000

MLA-C02 is three-quarters an operations exam. Reading covers the vocabulary; only production experience covers the judgement. If you have deployed a model, watched it drift and debugged why serving predictions differed from validation, you are most of the way there. If you have not, budget for building rather than reading.

Start with the official material

The exam guide. Four domains at near-equal weight: data preparation (28%), model and foundation model development (24%), deployment and orchestration (24%), and operating, monitoring and securing (24%). Only a quarter is model building. Let that shape your study.

N10-009 study resources

Sun, 20 Sep 2026 00:00:00 +0000

Network+ rewards building a network, breaking it and fixing it more than any other activity. Troubleshooting is the heaviest domain at 24%, and troubleshooting is not a thing you can read. CompTIA’s official objectives plus a home lab, virtual or physical, is the combination that passes this exam.

Start with the official material

The exam objectives PDF. CompTIA publishes a detailed objectives document listing every acronym and concept that can appear. It is exhaustive and free, and it is the definitive scope. Download it first.

SC-500 study resources

Sun, 20 Sep 2026 00:00:00 +0000

SC-500 punishes reading-only preparation more than most Microsoft exams. Its questions describe security requirements and ask which control satisfies them, which is a judgement you build by configuring things rather than by reading about them. Microsoft Learn plus a free Azure subscription and deliberate lab work is the combination that passes it.

Start with the official material

The skills outline. Four areas at roughly equal weight: identity, access and governance; securing storage, databases and networking; securing compute; and managing and monitoring security posture. No single area dominates, so you cannot afford a blind spot.

Spark Developer Associate resources

Sun, 20 Sep 2026 00:00:00 +0000

This is the most code-centric of the Databricks exams. DataFrame API work carries 30%, architecture 20% and Spark SQL 20%. The fastest preparation is writing Spark and reading query plans, not watching someone else do it. The official Spark documentation is excellent and free.

Start with the official material

The exam guide and its weights: DataFrame and DataSet API applications 30%, architecture and components 20%, Spark SQL 20%, troubleshooting and tuning 10%, structured streaming 10%, Spark Connect 5%, pandas API on Spark 5%.

200-901 cheat sheet: one-page reference

Fri, 09 Oct 2026 00:00:00 +0000

This is a one-page 200-901 reference for the night before: the exam format, HTTP status codes, authentication, Python requests, Git, Docker, bash, the Cisco platforms and the common ports. It does not replace the domain guides; it is the list you reread to check what you still remember.

Exam facts


Duration	120 minutes
Questions and passing score	Not published
Price	300 USD at time of writing
Languages	English, Japanese
Validity	3 years

Domains: Software Development and Design 15%, APIs 20%, Cisco Platforms and Development 15%, Application Deployment and Security 15%, Infrastructure and Automation 20%, Network Fundamentals 15%.

AB-100 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for AB-100. Deploy is the largest domain at 40–45%; Plan and Design are 25–30% each.

Exam facts


Passing score	700 / 1000
Duration and number of questions	Not published on the exam page
Price	Set per country or region; shown at booking
Languages	13
Renewal	Free, annually
Also required for the credential	One qualifying associate certification
Outline	Skills measured as of October 14, 2026

Skill areas

Area	Weight
Deploy AI-powered business solutions	40–45%
Plan AI-powered business solutions	25–30%
Design AI-powered business solutions	25–30%

Requirement to answer

Requirement	Answer
Fastest, cheapest start	Prebuilt agent, then extend
Users live in Microsoft 365	Extend Microsoft 365 Copilot
Agent acts on events, no user trigger	Autonomous agent, with limits and approvals
Agent answers a user’s question	Prompt-and-response agent
Defined steps on request	Task agent
Handle unmatched conversation	Fallback topic
Standard interface to tools	Model Context Protocol
Agents talk to agents across platforms	Agent2Agent
Application with no API	Computer use
Cheaper model for simple requests	Model router
Narrow, high-volume task, low cost	Small language model
Wrong answers, normal latency	Fix grounding data
Same promotion for many objects	Versioned, coordinated ALM
Prevent users seeing restricted content	Permissions on grounding data
Who changed the model or data	Audit trail
Region and data movement	Residency and movement validation

Agent types

Type	Trigger	Control to design
Prompt and response	User question	Grounding and answer quality
Task agent	User or process request	Defined steps, error handling
Autonomous agent	Event or schedule	Limits, approvals, escalation, audit

Platforms at a glance

Platform	Think of it as
Microsoft 365 Copilot	Where employees already work; extend it with agents
Copilot Studio	Low-code agents, topics, flows, connectors, MCP
Microsoft Foundry	Models, custom models and the Agents Service
Dynamics 365 apps	Built-in copilots and agents for finance, supply chain, sales and service
Power Platform	Apps and flows with AI features, governed by environments

Grounding data properties

Accuracy, relevance, timeliness, cleanliness, availability.

AB-730 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

This one-page AB-730 cheat sheet collects the four skill areas, the decisions the exam keeps testing, and the look-alike features that cause lost marks. It follows the outline in force from October 20, 2026. Read the domain guides first, then use this page for the last revision pass.

The exam in one table

Item	Value
Duration	45 minutes
Passing score	700 / 1000
Questions	Not published
Fundamentals	25–30%
Prompts and chats	20–25%
Content and collaboration	20–25%
Agents	20–25%

Which tool?

Situation	Choose
Ask or draft by conversation	Copilot Chat
Edit inside a document	Copilot in that app
Repeatable specialist task	Agent
Multi-step job you delegate and review	Copilot Cowork

Prompt actions

Verb	Feature
Reuse	Save
Run regularly	Schedule
Give to others	Share
Improve the answer	Refine
Keep preferences	Memory and instructions

Chats

Find, rename and delete chats. Group related material in a notebook and share it. Filter chats by agent.

AB-731 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for AB-731: three skill areas, the vocabulary, a product-matching table and the governance checklist. Business value and Microsoft AI apps carry about three quarters of the marks.

Exam facts


Duration	45 minutes
Passing score	700 / 1000
Weights	Business value 35–40%; Microsoft AI apps and services 35–40%; Adoption 20–25%
Outline	July 22, 2026
Languages	13

Vocabulary

Term	One line
Generative AI	Creates new content
Machine learning	Predicts or classifies from data
Pretrained model	General model, ready to use
Fine-tuned model	Pretrained model trained further on your examples
Token	Unit of text that drives usage cost
Prompt engineering	Better instructions, better output
Grounding	Tying answers to your data
RAG	Retrieve relevant content, then generate
Fabrication	Confident but invented answer

Which Microsoft product

Need	Answer
Secure chat for work accounts	Microsoft 365 Copilot Chat
Help inside Word, Excel, PowerPoint, Outlook, Teams	Copilot in the apps
Answers from my emails, meetings and files	Microsoft 365 Copilot with Microsoft Graph
Research report across web and work data	Researcher
Analyse spreadsheets and data files	Analyst
Low-code agent for a team	Copilot Studio
Extend Copilot with your data or actions	Microsoft 365 Copilot extensibility
Custom app, model choice, scale	Microsoft Foundry
Search and retrieve company content	Azure AI Search
Analyse images and video	Azure Vision in Foundry Tools

Order of preference: buy, extend, build.

AZ-305 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for AZ-305. Infrastructure is the largest domain; business continuity is the smallest. Everything below is a decision shortcut, so check the requirement before using one.

Exam facts


Passing score	700 / 1000
Number of questions	Not published
Duration	Not published on the exam page; confirm when booking
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, annually
Prerequisite for the certification	Azure Administrator Associate (AZ-104)

Domains

Area	Weight
Design infrastructure solutions	30–35%
Design identity, governance, and monitoring solutions	25–30%
Design data storage solutions	20–25%
Design business continuity solutions	15–20%

Identity, governance and monitoring

Requirement	Answer
Workload needs Azure access, no secret to manage	Managed identity
Store secrets, keys, certificates	Key Vault
Partners sign in with their own accounts	B2B guests
Customers sign up with social accounts	External identity service for customers
Publish an on-premises web app without a VPN	Entra application proxy
Just-in-time admin rights	Privileged Identity Management
Periodic recertification of access	Access reviews
Controls inherited by many subscriptions	Management group + Azure Policy
Block non-compliant deployments	Policy, Deny effect
Query logs across resources	Log Analytics workspace
Stream logs to a SIEM	Diagnostic setting to Event Hubs

Data

Requirement	Answer
SQL Server lift and shift, full compatibility	SQL Managed Instance
New app, managed relational	Azure SQL Database
Many databases, uneven peaks	Elastic pool
Intermittent load, pay less	Serverless compute tier
Global, low-latency NoSQL	Cosmos DB
Raw analytics data with folders	Data Lake Storage Gen2
Hybrid file server	Azure Files + Azure File Sync
Rarely read, long retention	Archive tier + immutability policy
Scheduled data movement	Data Factory
Streaming ingestion	Event Hubs

Redundancy and recovery

Failure to survive	Choice
Disk or rack	LRS
Datacenter	ZRS or availability zones
Region	GRS or GZRS
Read secondary during outage	RA-GRS or RA-GZRS
Regional SQL failover, same endpoint	Failover group
VM regional DR	Site Recovery
Restore points for VMs and files	Recovery Services vault
Restore points for disks and blobs	Backup vault
Accidental deletion or overwrite	Soft delete, versioning, backup

Replication copies mistakes too. Backup is what saves you from deletion and corruption.

AZ-400 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for AZ-400. Build and release pipelines are more than half the exam; instrumentation is the smallest area.

Exam facts


Duration	100 minutes (120 if the exam may contain labs)
Passing score	700 / 1000
Questions	Not published for this exam
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, annually

Skill areas

Area	Weight
Build and release pipelines	50-55%
Processes and communications	10-15%
Source control strategy	10-15%
Security and compliance plan	10-15%
Instrumentation strategy	5-10%

Requirement to answer

Requirement	Answer
Link commits and PRs to Azure Boards work items	Boards-GitHub integration, AB#id
Diagram in a wiki as text	Mermaid in Markdown
Call another system on an event	Webhook / service hook
Frequent deploys, strong tests, no long branches	Trunk-based + feature flags
Maintain several shipped versions	Release branches
Enforce reviews and builds before merge	Branch policy (Azure Repos) / branch protection (GitHub)
Large binaries in Git	Git LFS
Very large repository performance	Scalar, partial clone, sparse checkout
Undo a pushed commit safely	git revert
Find a deleted branch’s commits	git reflog
Secret committed	Rotate, then rewrite history
Same steps in many pipelines	Templates / reusable workflows
Approval that YAML edits cannot remove	Environment check
Private network build access, low upkeep	Scale set agents / self-hosted runners
Promote tested packages	Azure Artifacts views
Breaking change in SemVer	Increment MAJOR
Instant rollback, near-zero downtime	Blue-green / slot swap
Limited exposure first	Canary or ring
Switch a feature on without deploying	Feature flag, App Configuration
On-demand environments from approved templates	Azure Deployment Environments
Azure deployment without stored secret	Workload identity federation (OIDC)
Code on an Azure resource calls Azure	Managed identity
Per-run token inside a workflow	GITHUB_TOKEN
Central secret store	Azure Key Vault
File secrets in pipelines	Secure files
Vulnerable dependencies	Dependabot, dependency review
Code vulnerabilities	CodeQL
Pipeline and code findings in cloud security view	Defender for Cloud DevOps security
Request flow across services	Application Insights distributed tracing
Probe site from other regions	Availability test
AKS pod metrics	Container insights

Deployment strategies

Strategy	Key idea
Blue-green	Two environments, switch all traffic
Canary	Small share of traffic first
Ring	Gradual rings of users
Rolling	Batches of instances
Feature flag	Deploy off, release later
A/B	Compare variants

YAML keywords

Concept	Azure Pipelines	GitHub Actions
Trigger	`trigger`, `pr`	`on:`
Ordering	`dependsOn`	`needs:`
Condition	`condition:`	`if:`
Reuse	templates	reusable workflows, composite actions
Approval	environment checks	environment required reviewers

Common traps

Lead time versus cycle time: lead time starts at the request.
Managed identities cannot be attached to GitHub-hosted runners; use federation.
A custom condition replaces the default success check.
Deleting a secret from the latest commit leaves it in history.
Log alerts need a Log Analytics workspace; platform metrics do not.

More detail: exam format and study plan. Check the glossary for unfamiliar terms.

Azure administrator and architect path

Fri, 09 Oct 2026 00:00:00 +0000

The Azure route runs AZ-900 (optional), then AZ-104, then either AZ-305 for the architect credential or AZ-400 for DevOps. AZ-104 is the one you cannot skip: Microsoft names it as the prerequisite for the Azure Solutions Architect Expert certification.

The four exams

Exam	Certification	Take it if
AZ-900	Azure Fundamentals	You are new to cloud, or you work next to it
AZ-104	Azure Administrator Associate	You run Azure day to day
AZ-305	Azure Solutions Architect Expert	You design solutions and choose between services
AZ-400	DevOps Engineer Expert	You build pipelines, source control strategy and release processes

The prerequisite rule

Microsoft’s certification pages say that the Azure Solutions Architect Expert and DevOps Engineer Expert credentials need the Azure Administrator Associate certification first. The DevOps page also still mentions the Azure Developer Associate as an alternative, but Microsoft marks that certification as retired, so plan on AZ-104. The AZ-104 vs AZ-305 and AZ-400 vs AZ-104 guides go into detail.

CCNA 200-301 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for CCNA 200-301 on the v2.0 outline. Infrastructure and switching are half the exam; AI and operations is the smallest domain.

Exam facts


Duration	120 minutes
Questions / passing score	Not published
Price	300 USD at time of writing
Languages	English, Japanese
Validity	3 years

Domains

Domain	Weight
1.0 Network Infrastructure and Connectivity	25%
2.0 Switching and Network Access	25%
3.0 IP Routing	20%
4.0 Network Services and Security	20%
5.0 AI, and Network Operations and Management	10%

Subnetting

Prefix	Block	Hosts
/24	256	254
/25	128	126
/26	64	62
/27	32	30
/28	16	14
/29	8	6
/30	4	2

Usable hosts = 2^(host bits) - 2. Wildcard mask = 255.255.255.255 minus the subnet mask. A 169.254.x.x address means no DHCP reply.

DP-600 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for DP-600. Prepare data is the largest domain (45–50%); semantic models and maintaining the solution share the rest. This sheet follows the outline that applies from October 19, 2026.

Exam facts


Duration	100 minutes
Passing score	700 / 1000
Questions	Not published
Domains	Maintain 25–30%, Prepare data 45–50%, Semantic models 25–30%
Renewal	12 months, free online assessment

Which store

Item	Language	Use for
Lakehouse	Spark; read-only T-SQL endpoint	Files, Delta tables, notebooks
Warehouse	T-SQL, full DML	Relational, transactions, procedures
Eventhouse	KQL	Streaming, time-series

Getting data in

Need	Use
No copy, data stays where it is	Shortcut
Near-real-time replica of a database	Mirroring
Scheduled bulk load	Pipeline copy activity
Low-code Power Query shaping	Dataflow Gen2
Code transformation	Notebook
Load files into a warehouse	`COPY INTO`
Discover items / streams	OneLake catalog / Real-Time hub

Transform patterns

Latest row per key: ROW_NUMBER() OVER (PARTITION BY k ORDER BY d DESC), keep 1.
Replace nulls: COALESCE(col, 0).
No matching rows: left join, then WHERE right.key IS NULL.
Flatten snowflake: join to one dimension.
Filter rows: WHERE; filter groups: HAVING.

KQL and DAX syntax

Events | where Time > ago(1h) | summarize count() by Device

EVALUATE SUMMARIZECOLUMNS(Product[Category], "Sales", [Total Sales])

DAX: VAR ... RETURN for reuse; CALCULATE changes filter context; iterators (SUMX) go row by row; OFFSET and WINDOW are relative to the current row.

DP-700 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for DP-700. The three skill areas carry equal weight, so revise all three.

Exam facts


Duration	100 minutes
Passing score	700 / 1000
Questions	Not published
Price	Set per country or region; shown at booking
Languages	10
Renewal	Annual, free

Skill areas

Area	Weight
Implement and manage an analytics solution	30–35%
Ingest and transform data	30–35%
Monitor and optimize an analytics solution	30–35%

Requirement to answer

Requirement	Answer
Low-code transformation	Dataflow Gen2
Code at scale, libraries	Notebook
Orchestrate steps, retries, alerts	Pipeline
Run when a file arrives	Event-based trigger
Fixed cadence	Schedule
Different lakehouse per stage	Deployment rule
Version control	Git integration
Shared libraries for notebooks	Environment
Filter rows per user	Row-level security policy
Hide values, keep rows	Dynamic data masking
Access to one folder	OneLake security role
Organisation-approved content	Certified endorsement
Who did what	Audit log in Purview
Data without copying	OneLake shortcut
Keep a database in sync	Mirroring
Only changed rows	Incremental load with a watermark
Latest row per key	row_number over a window, keep 1
Sub-second time series	Eventstream, Eventhouse, KQL
Complex logic on streams	Spark structured streaming
Faster KQL on shortcut data	Query acceleration

Window types

Window	Remember
Tumbling	Fixed, no overlap; KQL `bin()`
Hopping	Fixed, overlapping by hop
Sliding	Output on enter and leave
Session	Ends after inactivity gap
Snapshot	Same timestamp

Symptom and cure

Symptom	Cure
Many small files	OPTIMIZE
Old unreferenced files	VACUUM (default retention 7 days)
Shuffle on small lookup	Broadcast join
Slow Copy of large table	Partition option, copy parallelism
Stale warehouse plans	Update statistics
Recent data slow in Eventhouse	Caching policy
Old data accumulating	Retention policy
Missing Eventhouse records	`.show ingestion failures`
Notify on failure	On fail path with Teams or email activity

Traps

OPTIMIZE and VACUUM are not the same.
Masking is not blocking. Rows and access are unchanged.
Workspace roles override narrow controls.
Promoted is not Certified.
A shortcut reads in place; a copy duplicates.
Do not load facts before dimensions.
collect() on big data causes driver out-of-memory.

Night-before checklist

Read the outline’s three headings, say the window types, and take one last practice test. Sleep. Details in the exam day guide.

DP-750 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for DP-750. Preparing data and deploying pipelines are the two largest areas, at 30–35% each.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Questions	Not published
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, every 12 months

Skill areas

Area	Weight
Prepare and process data	30–35%
Deploy and maintain data pipelines and workloads	30–35%
Set up and configure an Azure Databricks environment	15–20%
Secure and govern Unity Catalog objects	15–20%

Requirement to answer

Requirement	Answer
Scheduled unattended run, pay only for the run	Job compute or serverless jobs
Many concurrent SQL and BI users	SQL warehouse
Faster cluster start	Pool, or serverless
Stop paying for idle interactive compute	Auto-termination
Isolate prod from dev	Separate catalogs
Files must survive DROP TABLE	External table
Query a database in place	Connection plus foreign catalog
Read a table	SELECT plus USE CATALOG plus USE SCHEMA
Hide rows by user or group	Row filter
Mask values by user or group	Column mask
Rule driven by tags across many tables	ABAC
Secret without hard-coding	Key Vault-backed secret scope
Storage access without a stored key	Managed identity
Share data with an outside party	Delta Sharing (now OpenSharing)
Who uses this column?	Lineage in Catalog Explorer
Keep history of changes per row	SCD type 2 or temporal table
Upsert from a CDC feed	MERGE
High-cardinality or changing filters	Liquid clustering
New files only, scalable, schema evolution	Auto Loader
SQL-only incremental file load	COPY INTO
Managed ingestion from supported sources	Lakeflow Connect
Continuous events from Event Hubs	Structured Streaming with a checkpoint
Bad rows dropped, pipeline continues	Expectation with drop action
Stop when bad data appears	Expectation with fail action
Reject mismatched schema	Schema enforcement
Accept a new column	Schema evolution

Jobs and deployment

Need	Use
Order of tasks	Task dependencies
Start on new files	File arrival trigger
Transient failure	Automatic retries
Rerun only failed part	Repair run
Cancel a run	Stop
Package and deploy as code	Bundle, `databricks bundle deploy -t <target>`
Change history and review	Git folders, pull requests

Tuning and monitoring

Symptom or goal	Tool or command
Small files slow reads	OPTIMIZE
Old unreferenced files cost storage	VACUUM (default 7-day retention)
One task far slower than others	Skew; Spark UI and query profile
Disk spill	More memory per task
Central logs and alerts	Log Analytics and Azure Monitor

Traps

SELECT alone is not enough to read a table.
OPTIMIZE compacts; VACUUM deletes. They are not interchangeable.
Managed tables lose their data on DROP; external ones do not.
Names change: Lakeflow Spark Declarative Pipelines, Declarative Automation Bundles.

Take the practice test next, or revisit the exam tips.

DP-800 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for DP-800. Database development and security/deployment are the two largest areas at 35–40% each; AI capabilities is 25–30%.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Questions	Not published
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, annually
Outline	Skills measured as of October 19, 2026

Skill areas

Area	Weight
Design and develop database solutions	35–40%
Secure, optimize, and deploy database solutions	35–40%
Implement AI capabilities in database solutions	25–30%

Pick the table type

Need	Use
Data as of a past time	System-versioned temporal
Tamper-evidence	Ledger
Relationships traversed with MATCH	Graph
Very high concurrency, low latency	Memory-optimized
Large scans and aggregates	Columnstore index
Fast removal of old data	Partitioning with switch-out
Shared counter across tables	Sequence

T-SQL function families

Purpose	Functions
JSON	OPENJSON, JSON_VALUE, JSON_OBJECT, JSON_ARRAY, JSON_ARRAYAGG, JSON_CONTAINS
Regular expressions	REGEXP_LIKE, REGEXP_SUBSTR, REGEXP_REPLACE, REGEXP_COUNT, REGEXP_INSTR, REGEXP_MATCHES, REGEXP_SPLIT_TO_TABLE
Fuzzy matching	EDIT_DISTANCE, EDIT_DISTANCE_SIMILARITY, JARO_WINKLER_DISTANCE
Vectors	VECTOR_DISTANCE, VECTOR_NORMALIZE, VECTORPROPERTY, VECTOR_SEARCH

Security: which feature

Requirement	Feature
Mask for most users	Dynamic Data Masking (UNMASK to exempt)
Plaintext never on server	Always Encrypted
Per-user rows	Row-Level Security, filter plus block predicate
No passwords in config	Entra authentication, managed identity
Record activity	Auditing

Performance

Symptom	Tool or fix
New plan slower	Query Store, force plan
Top consumers in Azure SQL	Query Performance Insight
Deadlock	Same access order, short transactions
Readers block writers	Row versioning isolation

Reacting to changes

Option	Gives you
Change Tracking	Which rows changed
CDC	Before and after values
Change event streaming	Changes as events
Functions SQL trigger binding	Code run on changed rows

Search choices

Need	Use
Exact words	Full-text
Meaning	Vector
Both	Hybrid, merged with RRF: sum of 1 / (k + rank)
Large table, speed over perfect recall	ANN with vector index
Small table, exact results	KNN, VECTOR_DISTANCE

RAG in five steps

Embed the question, retrieve rows, convert to JSON, send with sp_invoke_external_rest_endpoint, extract the answer.

PL-300 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for PL-300. Prepare, model and visualise are equal in weight; manage and secure is the smallest.

Exam facts


Duration	100 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, every 12 months

Skill areas

Area	Weight
Prepare the data	25–30%
Model the data	25–30%
Visualize and analyze the data	25–30%
Manage and secure Power BI	15–20%

Requirement to answer

Requirement	Answer
Stack two tables with the same columns	Append Queries
Add columns from another table by key	Merge Queries
Months as columns, need one row per month	Unpivot Other Columns
Derived query that follows the original	Reference
Independent copy of a query	Duplicate
Switch source server without editing queries	Parameter
Profile all rows, not the first 1000	Column profiling on entire data set
Very large source, live data	DirectQuery
Second date column on the fact table	Inactive relationship + USERELATIONSHIP
Ratio that totals correctly	Measure with DIVIDE
Percent of total	CALCULATE with ALL / REMOVEFILTERS
Year to date, last year	TOTALYTD, SAMEPERIODLASTYEAR on a marked date table
Balance that must not add across time	Semi-additive: LASTDATE / CLOSINGBALANCE
Same time variants across many measures	Calculation group
Find the slow visual	Performance Analyzer
Smaller model	Remove columns, reduce granularity, star schema
Right-click to a filtered detail page	Drillthrough
Small chart on hover	Tooltip page
Toggle between visuals	Bookmarks + buttons + Selection pane
Same slicer on several pages	Sync slicers
Pixel-perfect printable output	Paginated report
Phone-friendly	Phone (mobile) layout
Why did the value rise	Analyze: Explain the increase
Read-only content for many consumers	App with audiences
On-premises source refresh	On-premises data gateway
Alert when a tile passes a value	Data alert on a dashboard tile
Only authorised reviewers mark trusted	Certified endorsement
Each user sees own rows	Dynamic RLS with USERPRINCIPALNAME

DAX patterns

Pattern	Expression
Percent of total	`DIVIDE([Sales], CALCULATE([Sales], ALL('Product')))`
Last year	`CALCULATE([Sales], SAMEPERIODLASTYEAR('Date'[Date]))`
Year to date	`TOTALYTD([Sales], 'Date'[Date])`
Use inactive relationship	`CALCULATE([Sales], USERELATIONSHIP(Sales[ShipDate], 'Date'[Date]))`
Closing balance	`CALCULATE(SUM(Stock[Units]), LASTDATE('Date'[Date]))`
Dynamic RLS filter	`[Email] = USERPRINCIPALNAME()`

Workspace roles

Role	Remember
Admin	Everything
Member	Edits, publishes and updates the app, adds people with equal or lower roles
Contributor	Creates and edits content, cannot manage access
Viewer	Views only; RLS applies

Common traps

Merge adds columns; append adds rows.
Only one active relationship between two tables.
Time intelligence needs a contiguous, marked date table.
Calculated columns are row-by-row at refresh; measures respond to filter context.
RLS roles need members added in the service; Admin, Member and Contributor bypass them.
Promote is open to writers; certify is for authorised reviewers.
A cloud source usually needs no gateway.

See the exam tips and the study plan for the longer version.

SC-200 cheat sheet: one-page reference

Fri, 09 Oct 2026 00:00:00 +0000

This is a one-page SC-200 reference for the final days before the exam: the facts, the three domains, and the comparisons that questions are built around. It is a revision aid, not a replacement for the study plan.

Exam facts


Duration	100 minutes
Passing score	700 / 1000 (scaled)
Questions	Not published
Languages	10
Renewal	Free, every 12 months
English update	October 21, 2026

Domains

Domain	Weight
Manage a security operations environment	40–45%
Respond to security incidents	35–40%
Perform threat hunting	20–25%

Sentinel automation

Need	Use
Assign, tag, set status, close noisy incidents	Automation rule
Call other systems or run multi-step workflows	Playbook (Logic Apps)
Fastest detection on a simple query	NRT analytics rule
Joins or longer lookback	Scheduled analytics rule
Unusual behaviour via ML	Anomaly rule

Sentinel roles

Role	Can
Reader	View
Responder	Manage incidents
Contributor	Create and edit content
Automation Contributor	Service role for playbooks

Data in

Source	Method
Windows security events	AMA connector plus data collection rule
Windows events via forwarders	Windows Event Forwarding
Linux and appliances	Syslog or CEF via AMA
Azure activity	Azure Policy and diagnostic settings
No built-in table	Custom log table
Rarely queried bulk data	Data lake tier

Defender for Endpoint response

Action	Purpose
Isolate device	Contain
Live response	Interactive shell, collect files
Investigation package	Collect a snapshot of evidence
Device timeline	Reconstruct what happened
ASR rules	Prevent risky behaviours
Automation level (per device group)	Approval or full automatic remediation

Microsoft 365 investigation

Tool	Finds
Purview Audit	User and admin activity
eDiscovery content search	Items by content
Graph activity logs	Graph API calls

Hunting tables (Defender XDR)

DeviceProcessEvents (processes), DeviceNetworkEvents (connections), DeviceLogonEvents (logons), DeviceFileEvents (files), EmailEvents (mail), IdentityLogonEvents (identity logons), CloudAppEvents (cloud apps). Defender uses Timestamp; Sentinel uses TimeGenerated.

SC-300 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for SC-300. Authentication and access management is the largest area; the other three are about equal.

Exam facts


Duration	100 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, every 12 months
Outline	As of October 28, 2026 (English)

Skill areas

Area	Weight
Implement and manage user identities	20–25%
Implement authentication and access management	25–30%
Plan and implement workload identities	20–25%
Plan and automate identity governance	20–25%

Requirement to feature

Requirement	Answer
Admin rights for a subset of users only	Role scoped to an administrative unit
Membership follows an attribute	Dynamic group (needs P1 or higher)
Link two tenants of one organisation	Cross-tenant synchronization
Skip repeat MFA for partner guests	Inbound trust of MFA claims in cross-tenant access settings
Several forests, minimal infrastructure	Entra Cloud Sync
Sign-in survives an on-premises outage	Password hash sync
Passwords checked on-premises	Pass-through authentication
Bootstrap passwordless	Temporary Access Pass
Gradual push to Authenticator	Registration campaign
SSPR must change on-premises password	Password writeback
Try a policy without enforcing	Report-only mode, What If
Step-up for one action	Authentication context
Revoke access in near real time	Continuous access evaluation
Compromised credentials	User risk policy: password change
Suspicious sign-in	Sign-in risk policy: MFA
Private apps, no VPN	Entra Private Access
Secret-free access from an Azure resource	Managed identity
Secret-free access from a pipeline	Federated credential
Daemon calls an API	Application permission plus admin consent
Users can’t consent	Admin consent workflow
On-premises web app, Conditional Access in front	Application Proxy
Block downloads in a SaaS session	Session policy via Conditional Access app control
Time-limited partner access	Access package with expiry
Recurring check of access	Access review with auto-apply
Standing admin access removed	PIM eligible assignment
Lockout recovery	Two cloud-only emergency accounts
Query sign-ins	`SigninLogs` in Log Analytics

Conditional Access rules

All matching policies apply; block beats grant.
Evaluated after first-factor authentication.
Test in report-only mode and with What If.
Exclude emergency access accounts.
Security defaults and Conditional Access cannot be combined.

Quick distinctions

Pair	Difference
Delegated vs application	A user is present vs no user, admin consent
Eligible vs active	Needs activation vs in force now
System vs user-assigned identity	One resource vs shared and reusable
Joined vs registered vs hybrid	Corporate cloud-only; personal; corporate and in AD
App registration vs enterprise application	The app’s definition vs the service principal in your tenant
Log Analytics vs storage vs Event Hubs	Query; cheap archive; stream

Common traps

Assuming Global Administrator reads custom security attributes: it needs the attribute roles.
Forgetting that application permissions always require admin consent.
Hiding an app from My Apps and thinking access is blocked: use assignment required.
Treating a report-only policy as enforced.
Using a one-time review where a recurring one is asked.

See the study plan and the practice test to apply this.

SC-401 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for SC-401. The three domains each carry 30–35%, so revise all of them.

Exam facts


Duration	100 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Languages	7
Renewal	Free, every 12 months
Questions	Not published

Skill areas

Area	Weight
Implement information protection	30–35%
Implement data loss prevention and retention	30–35%
Manage risks, alerts, and activities	30–35%

Requirement to answer

Requirement	Answer
Match actual values from a customer table	Exact data match
Match documents built from a template	Document fingerprinting
Recognise content by meaning	Trainable classifier
Detect text in images	OCR
Encrypt a document and travel with it	Sensitivity label with encryption
Label files already stored in SharePoint	Service-side auto-labeling policy
Test before enforcing	Simulation mode
Label has no effect for users	Publish it in a label policy
Label a Teams team or site	Label scope includes groups and sites
Bulk-label on-premises shares	Information Protection scanner, discovery first
Encrypted external email	Message Encryption
Expiry and revocation of email	Advanced Message Encryption
Block external sharing of card numbers	DLP policy with block
Let user proceed with a reason	Block with override
Control USB, print, paste	Endpoint DLP on onboarded devices
Monitor Chrome or Firefox	Purview browser extension
DLP rule that reacts to risky users	Adaptive Protection risk level
One rule for a whole location	Retention policy
Different rule per item, records, disposition review	Retention label
Retention follows department changes	Adaptive scope
Which retention rules apply here	Policy lookup
Detect departing-employee exfiltration	Data theft by departing users + HR connector
Capture device activity of a suspect	Forensic evidence
Hide usernames from analysts	Pseudonymisation
Longer audit record retention	Audit retention policy (right licence)
Label and DLP activity history	Activity explorer
Search, hold, export for a legal matter	eDiscovery
See and protect AI data use	DSPM for AI (auditing on)

Precedence reminders

DLP across policies: the more restrictive action applies.
Retention: retention wins over deletion; the longest retention wins; explicit wins over implicit.

Traps

A retention label is not a sensitivity label.
Sensitivity labels protect; retention labels keep or delete.
Insider Risk Management detects; DLP blocks.
No onboarded device means no Endpoint DLP.
No auditing means little AI visibility.

Study more with the guides or the practice test.

SC-900 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for SC-900. Security solutions is the largest area; concepts is the smallest, but its vocabulary supports the rest.

Exam facts


Duration	45 minutes
Passing score	700 / 1000
Questions	Not published
Renewal	None
Weights	Concepts 10–15%, Entra 25–30%, Security 35–40%, Compliance 20–25%

Concepts

Zero Trust: verify explicitly, least-privilege access, assume breach.
Defense in depth: layered controls; one failure does not expose everything.
Shared responsibility: the provider secures the facility; the customer always owns data and identities.
Encryption reversible with a key; hashing one-way.
Authentication proves who; authorization decides what; federation links identity providers.

Microsoft Entra

Need	Answer
Conditions on sign-in	Conditional Access
Time-limited role elevation	Privileged Identity Management
Detect risky sign-ins	ID Protection
Recertify access periodically	Access reviews
Manage joiners, movers, leavers	ID Governance
One identity on-premises and cloud	Hybrid identity
Identity for apps and services	Workload identity
Users reset own password	Self-service password reset

MFA = two of: know, have, are. Passwordless: Authenticator, FIDO2 key, Windows Hello for Business.

SY0-801 cheat sheet

Fri, 09 Oct 2026 00:00:00 +0000

Last-minute reference for SY0-801. Security Operations and Threats are the largest domains; program management is the smallest.

Exam facts


Questions	Maximum 90, multiple choice and performance-based
Duration	90 minutes
Passing score	750 on a scale of 100-900
Launch	Expected on or around November 17, 2026
Price	Not published at time of writing
Recommended experience	Security administrator, two years hands-on

Domains

Domain	Weight
Security Operations	27%
Threats, Vulnerabilities, and Attacks	24%
Security Architecture	19%
General Security Concepts	16%
Security Program Management and Oversight	14%

Control types

Type	Think
Preventive	Stops it
Deterring	Discourages it
Detective	Notices it
Corrective	Fixes it
Compensating	Stand-in for a missing control
Directive	Instructs people

Clue to answer

Clue	Answer
Near-real-time certificate status	OCSP
One password, many accounts	Password spraying
QR code leads to fake login	Quishing
Hidden instructions in content a model reads	Prompt injection
Two locations, impossible timing	Impossible travel
Time-limited admin rights	Just-in-time access
Maximum acceptable data loss	RPO
Maximum acceptable downtime	RTO
Real value in vault, reference elsewhere	Tokenization
Who handled the evidence	Chain of custody
Insurance	Transfer risk
Right to inspect vendor	Right to audit
Specific project work	SOW

Formulas

SLE = asset value x exposure factor
ALE = SLE x ARO

Email authentication

SPF: allowed senders. DKIM: signature. DMARC: policy and reporting. BIMI: brand logo.

Is AZ-802 worth it? Who should take it

Tue, 29 Sep 2026 00:00:00 +0000

AZ-802 is worth it if you administer Windows Server, on-premises or in Azure, and want a credential that says so. It is now the only exam that earns the Microsoft Certified: Windows Server Administrator Associate certification, it covers the work of that role closely, and one exam is cheaper and quicker than the two it replaces. It is less worth it if you already hold the Windows Server Hybrid Administrator certification, or if your work is mostly in Azure rather than on Windows Server.

AB-900 cheat sheet

Mon, 28 Sep 2026 00:00:00 +0000

Last-minute reference for AB-900. Data protection and governance is the largest domain; Copilot and agents is the smallest.

Exam facts


Duration	45 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Level	Fundamentals; does not expire
Outline version	As of October 14, 2026

Skill areas

Area	Weight
Understand data protection and governance tasks for Microsoft 365 and Copilot	35–40%
Identify the core features and objects of Microsoft 365 services	30–35%
Perform basic administrative tasks for Copilot and agents	25–30%

Requirement to tool

Requirement	Answer
Add a custom domain	Microsoft 365 admin center
Licences follow team membership	Group-based licensing
One inbox, several users	Shared mailbox
Email to every member’s own inbox	Distribution group
Read-only access to a site	Visitors group
MFA only outside the office	Conditional access with named locations
Why was a sign-in blocked?	Entra sign-in logs
Who changed a role assignment?	Entra audit logs
Admin roles only when needed	Privileged Identity Management
SSO for a gallery SaaS app	Enterprise applications
One incident across email, identity and endpoint	Defender XDR
Encrypt a file wherever it goes	Sensitivity label with encryption
Block sensitive data being sent out	DLP policy
Keep for seven years, then delete	Retention policy or label
Where is sensitive data stored?	Data Explorer
What did users do with it?	Activity explorer
Departing user exfiltrating data	Insider Risk Management
Harassment in chats	Communication Compliance
AI use and sensitive prompts	DSPM for AI
Regulatory score and actions	Compliance Manager
Find emails for a legal case	eDiscovery Content search
Sites shared too broadly	Data access governance report
Limit a site to one group	Restricted access control (SAM)
Daily Copilot use in apps	Per-user monthly licence
Occasional or agent-only use	Pay-as-you-go billing policy
Research report from many sources	Researcher
Explain trends in spreadsheets	Analyst
Copilot adoption numbers	Usage reports, Copilot Analytics
Agent for one group only	Agent settings, Microsoft 365 admin center
Copilot Studio environments	Power Platform admin center

Zero Trust

Verify explicitly. Use least-privilege access. Assume breach.

AI-300 study resources

Mon, 28 Sep 2026 00:00:00 +0000

The best AI-300 resources are free: Microsoft’s skills outline, the official practice assessment, the Azure Machine Learning and Microsoft Foundry documentation, and an Azure subscription you build things in. Add realistic practice questions to find your weak domain. The one thing to be careful with is older material: much of what is labelled AI-300 online was written for DP-100 and covers only the machine learning half.

Start with the official material

The skills outline. Five domains, from 10–15% to 25–30%, split between Azure Machine Learning and Foundry. It names every tool and technique the exam covers: MLflow, registries, endpoints, Bicep, GitHub Actions, provisioned throughput, evaluators, hybrid search and fine-tuning. Read it first and keep it open while you study.

AZ-104 cheat sheet

Mon, 28 Sep 2026 00:00:00 +0000

Last-minute reference for AZ-104. Identity and compute are the largest areas; monitoring and backup is the smallest.

Exam facts


Duration	100 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Languages	10
Renewal	Free, every 12 months

Skill areas

Area	Weight
Manage Azure identities and governance	20–25%
Deploy and manage Azure compute resources	20–25%
Implement and manage storage	15–20%
Implement and manage virtual networking	15–20%
Monitor and maintain Azure resources	10–15%

Requirement to answer

Requirement	Answer
Group membership follows a user attribute	Dynamic membership group
Manage resources but not grant access	Contributor
Grant access but not manage resources	User Access Administrator
Reset passwords, no resource access	Entra role, not an Azure role
Block resources without a tag	Azure Policy, Deny effect
Fix existing non-compliant resources	Remediation task
Prevent deletion, allow changes	CanNotDelete lock
One assignment for many subscriptions	Management group
Revoke a group of SAS tokens	Stored access policy
SAS without the account key	User delegation SAS
Survive a datacenter failure (storage)	ZRS or GZRS
Read during a regional outage	RA-GRS or RA-GZRS
Recover overwritten blobs	Versioning
Recover deleted blobs	Soft delete
Move blobs to cheaper tiers by age	Lifecycle management
Survive a datacenter failure (VMs)	Availability zones
Release with instant rollback on App Service	Deployment slot + swap
Simple short-lived container	Container Instances
RDP without public IPs	Azure Bastion
PaaS on a private IP, reachable on-premises	Private endpoint
Force traffic through an appliance	Route table, next hop virtual appliance
Silence alerts during maintenance	Alert processing rule
Who deleted a resource	Activity log
VM backups	Recovery Services vault
Disk and blob backups	Backup vault
Run VMs in another region after an outage	Site Recovery

Numbers worth knowing


Reserved addresses per subnet	5
Bastion subnet	Named AzureBastionSubnet, /26 or larger
NSG rule priorities	100–4096, lowest number first
Minimum tier days	Cool 30, cold 90, archive 180
App Service autoscale and slots	Standard tier or higher

Traps

Peering is not transitive. Spokes do not talk through a hub without a gateway or appliance.
Traffic must pass both NSGs when subnet and NIC each have one.
Tags do not inherit from resource groups without a policy.
Locks apply to Owners too. A ReadOnly lock can block key listing on a storage account.
Budgets alert; they do not stop resources.
Complete deployment mode deletes resources not in the template.
Entra roles and Azure roles are separate. Global Administrator does not mean subscription access.
Archive blobs must be rehydrated before they can be read.

Night-before checklist

Name the five areas and which are largest
Least-privilege role and scope for any requirement, cold
Service endpoint versus private endpoint, in one sentence each
Check ID and proctoring rules; see exam day

Take the 20-question practice test and check your weakest area.

AZ-900 cheat sheet

Mon, 28 Sep 2026 00:00:00 +0000

Last-minute reference for AZ-900. Architecture and services is the largest area; the look-alike tools decide most marks.

Exam facts


Duration	45 minutes (65 minutes seat time)
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Level	Fundamentals
Expires	Never; no renewal

Skill areas

Area	Weight
Describe Azure architecture and services	35–40%
Describe Azure management and governance	30–35%
Describe cloud concepts	25–30%

Shared responsibility

	IaaS	PaaS	SaaS
Data, devices, accounts	You	You	You
Applications	You	Shared	Microsoft
Operating system	You	Microsoft	Microsoft
Physical datacenter	Microsoft	Microsoft	Microsoft

Storage redundancy

Option	Survives
LRS	Disk or server failure
ZRS	Datacenter (zone) failure
GRS	Region failure
GZRS	Zone and region failure

Add RA- for read access to the secondary region. Tiers: hot, cool, cold, archive (offline, rehydrate first).

AI-200 cheat sheet

Sun, 27 Sep 2026 00:00:00 +0000

Last-minute reference for AI-200. The data domain is the largest; the other three are roughly a quarter each.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Replaces	AZ-204
Language focus	Python

Skill areas

Area	Weight
Develop AI solutions by using Azure data management services	25–30%
Develop containerized solutions on Azure	20–25%
Connect to and consume Azure services	20–25%
Secure, monitor, and troubleshoot Azure solutions	20–25%

Requirement to service

Requirement	Answer
Rebuild the image when the base image is patched	ACR task with base image trigger
Run exactly the tested image	Reference by digest
Secret in a container on App Service	Key Vault reference + managed identity
Canary release, instant rollback	Container Apps multiple revision mode
Scale a worker on queue length, to zero	KEDA Service Bus rule, min replicas 0
Existing Kubernetes manifests	AKS
Read your own writes, cheapest	Cosmos DB session consistency
React to new documents in Cosmos DB	Change feed processor
Vector search with SQL filters and joins	PostgreSQL + pgvector
Semantic cache, lowest latency	Azure Managed Redis vector index
One consumer per message, retries	Service Bus queue
Every subscriber gets a copy, with filters	Service Bus topic + subscriptions
Notify many handlers of an event	Event Grid
Ordered processing per entity	Service Bus sessions
No credential stored anywhere	Managed identity + DefaultAzureCredential
Same key, different value per environment	App Configuration labels
Follow one request across services	OpenTelemetry context propagation

Vector indexes

Service	Options
Cosmos DB	`flat` (exact, small dimensions), `quantizedFlat`, `diskANN` (large scale)
PostgreSQL	IVFFlat (needs data first), HNSW (works on empty tables, higher recall)
Redis	Search index with a vector field

pgvector operators: <-> Euclidean, <=> cosine, <#> negative inner product. The operator must match the index’s operator class.

SEE Part 1 study resources: free IRS material

Mon, 21 Sep 2026 00:00:00 +0000

The best study material for SEE Part 1 is free: the IRS’s own 2025 publications and form instructions, plus the official Candidate Information Bulletin from PSI. The exam tests the law as amended through December 31, 2025, and questions are about calendar year 2025, so the 2025 editions are the source that matches the exam most closely. Commercial courses are useful for practice questions, but everything they teach comes from these documents.

SEE Part 2 cheat sheet

Mon, 21 Sep 2026 00:00:00 +0000

A one-page reference for SEE Part 2, for the last week before the exam. Every figure is for tax year 2025, the year the current exam tests, taken from the 2025 IRS publications and form instructions.

Exam facts


Questions	100 (85 scored, 15 unscored)
Time	3.5 hours; sections 1–34, 35–67, 68–100, locked once left
Pass	500 scaled (200–800)
Fee	$317
Domains	Entities 30, Tax preparation 37, Specialized 18

Basis and distributions, in order

Partner’s basis: + contributions (adjusted basis), + income, + liability increase; − distributions, − losses, − liability decrease. Never below zero.
S shareholder losses: limited to stock basis + debt basis from the shareholder’s own loans.
C corporation distribution: dividend to extent of E&P → return of stock basis → capital gain.
Section 351: 80% control; gain recognised up to boot; liabilities over basis create gain; stock for services is compensation.

Entities


Two-member LLC, no election	Partnership, Form 1065
Single-member LLC, no election	Disregarded
S corporation	≤100 shareholders, one class of stock, no nonresident aliens
Form 2553	Within 2 months 15 days of the year’s start, or during the prior year
Corporate tax rate	21%
Dividends-received deduction	50% (<20% owned), 65% (20%+ owned)
Corporate charitable limit	10% of taxable income (modified)
Corporate estimated tax	Required if expected tax is $500+
Centralized audit opt-out	100 or fewer eligible partners; no partnership or trust partners

2025 business figures


Section 179	$2,500,000; phase-out above $4,000,000
Special depreciation	100% for property acquired and placed in service after Jan 19, 2025 (40% by election)
Listed property	>50% business use for 179 or special allowance
Standard mileage	70 cents per mile
Meals / gifts	50% / $25 per recipient
Start-up and organizational costs	$5,000 each, reduced above $50,000; rest over 180 months
QBI	Up to 20%; Form 8995 if taxable income ≤ $197,300 ($394,600 MFJ)
Form 1099-NEC	$600+ for 2025 payments ($2,000 only after 2025)
Form 8300	Cash over $10,000
SE tax	Net earnings of $400+
Corporate NOL	Forward indefinitely, 80% of taxable income; carryback only for farming and certain insurance losses
Like-kind	Real property only; identify in 45 days, receive in 180
Recovery periods	Residential rental 27.5 years, nonresidential 39

Specialized returns


Form 1041 filing	Gross income $600+ (estates and trusts); any taxable income (trusts)
Exemptions	Estate $600; simple trust $300; other trust $100
65-day election	Complex trusts and estates
Form 990	Receipts $200,000+ or assets $500,000+; 990-N if normally ≤ $50,000
Form 990-T	Unrelated business gross income $1,000+
SEP	Lesser of 25% of compensation or $70,000; compensation cap $350,000
SIMPLE / 401(k) deferral	$16,500 / $23,500
Rental allowance	$25,000, phased out between MAGI $100,000 and $150,000
Real estate professional	>50% of services and >750 hours in real property trades
Farmer estimated tax	Two-thirds farming: file and pay by March 2, 2026, or one payment by January 15, 2026

Due dates for 2025 calendar-year returns

Return	Due
Form 1065, 1120-S	March 16, 2026
Form 1120, 1041	April 15, 2026
Form 990 series	15th day of 5th month

Traps

Fair market value is not basis on contributions.
Liabilities move partner basis in both directions.
Gift limit ≠ meal limit.
2025, not 2026. The 1099-NEC threshold change is the classic example.
EXCEPT questions: find the false one.

Night-before checklist

The basis orderings above, cold
The 2025 figure table, cold
Section checkpoints: ~2:19 remaining after section 1, ~1:10 after section 2
ID and remote setup checked

Take the 20-question practice test once more, and reread the exam tips.

SEE Part 3 cheat sheet

Mon, 21 Sep 2026 00:00:00 +0000

Last-minute reference for SEE Part 3. Everything here is a rule to recall, not a calculation — which is exactly how Part 3 is tested.

Exam facts


Questions	100 (85 scored)
Time	3.5 hours, three locked sections
Pass mark	500 scaled (200–800)
Fee	$317

Domains

Domain	Scored questions
Practices and Procedures	26
Representation before the IRS	25
Specific Areas of Representation	20
Filing Process	14

Circular 230 in one table

Section	Rule
§10.20	Give the IRS requested records promptly unless privileged
§10.21	Tell the client about an error or omission and its consequences
§10.23	No unreasonable delay
§10.27	No unconscionable fees; contingent fees only for exams of original returns (or claims filed within 120 days of an exam notice), interest or penalty claims, and court cases
§10.28	Return client records promptly, even in a fee dispute
§10.29	Conflicts: written consent within 30 days, kept 36 months
§10.30	EAs may not say “certified” or imply IRS employment
§10.31	Never negotiate a client’s refund check
§10.50	Censure, suspension, disbarment, monetary penalty

CE: 72 hours per 3-year cycle; 16 a year minimum, 2 of them ethics.

AAIA cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AAIA. The exam rewards the answer an auditor could evidence, not the one that sounds most thorough.

Exam facts


Questions	90
Domains	3
Requires	CISA or an equivalent designation
Exam fee	459 USD member / 599 USD non-member, plus a 50 USD application fee

Domain weights

Domain	Weight
AI Operations	46%
AI Governance and Risk	33%
AI Auditing Tools and Techniques	21%

Operations is the largest, which surprises people expecting a governance-heavy exam.

AAIR cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AAIR. The exam is about running a programme: who decides, on what evidence, and what changes as a result.

Exam facts


Questions	90
Domains	3
Largest domain	42%
Exam fee	459 USD member / 599 USD non-member, plus application fee

Domain weights

Domain	Weight
AI Risk Program Management	42%
AI Risk Governance and Framework Integration	37%
AI Life Cycle Risk Management	21%

Programme plus governance is nearly 80% of the exam.

AAISM cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AAISM. The technical domain is the largest, but the reasoning stays managerial: which control removes the mechanism, and how you would evidence it operates.

Exam facts


Questions	90
Domains	3
Requires	CISM or CISSP
Exam fee	459 USD member / 599 USD non-member, plus application fee

Domain weights

Domain	Weight
AI Technologies and Controls	38%
AI Governance and Program Management	31%
AI Risk Management	31%

The five control principles

Internalise these and most of the technical domain answers itself.

AB-620 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AB-620. This is a design exam: it asks which construction satisfies a stated constraint.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Level	Associate
Platform	Copilot Studio

Skill areas

Area	Weight
Integrate and extend agents in Copilot Studio	40–45%
Plan and configure agent solutions	30–35%
Test and manage agents	20–25%

Requirement to construction

Requirement	Answer
Answers must track weekly document updates	Ground on the knowledge source the owners maintain
Agent must never advise on a topic	Out-of-scope topic with a scripted refusal
Must create a ticket in another system	Action or connector call, not a link
Must know the user’s department	Read the authenticated user’s profile
Answers must be verifiable	Enable citations
Secret for a partner API	Secure credential store on the connection
Value needed later in the conversation	Conversation-scoped variable
External users must not see pricing	Separate sources; ground externally on public content only
Change tested before all users get it	Publish to a test environment or limited channel
Record of what the agent said	Transcript retention with access controls

Traps

Dev works, production fails → connection references and environment variables not configured in the target environment. This is both the real-world failure and the exam’s favourite.
Many phrasings not recognised → add varied trigger phrases and use generative understanding, not a longer response.
Topic-scoped variables die with the topic. Use conversation scope to carry values.
Instructing the model to respect permissions is not a control.
Stale answers after a document update → the knowledge index has not refreshed; the model did not memorise anything.
Cost is driven by conversation or message volume, not authors or documents.

Night-before checklist

Three areas; integration is the biggest block
Action versus generated text: deterministic, permission-aware, auditable
Escalation transfers context, it does not apologise
Check ID and proctoring rules — see exam day

Take the 20-question practice test first.

Agentforce Specialist cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Agentforce Specialist exam. The section people lose marks on is Data 360, not prompt engineering.

Exam facts


Questions	60 scored
Duration	105 minutes
Passing score	72%
Price	200 USD

A published percentage, unusually — so practice scoring is directly meaningful. Aim comfortably above it.

Section weights

Section	Weight
AI Agents	35%
Prompt Engineering	20%
Data 360 Fundamentals	20%
Testing, Deployment, and Maintenance	10%
Governance and Observability	10%
Multi-Agent Orchestration	5%

Agent constructs

Construct	Purpose
Topic	What the agent handles; its classification description drives routing
Action	A deterministic, permission-aware operation — always preferred over generated text for a business operation
Scope	What is explicitly out of bounds, with a scripted refusal
Escalation	Transfer to a human with transcript and context preserved
Variable	Conversation-scoped to carry values between topics

Data 360 essentials

Identity resolution — match and merge records across systems into one profile
Harmonised data model — map source fields to a standard structure
Streaming vs batch — continuously changing behavioural data needs streaming
Why quality matters more here — the agent acts on and asserts the data to customers

Recurring right answers

Ground the template with merge fields from the related record, never typed values
Specify the output structure explicitly when summaries come back inconsistent
Require human confirmation before a refund or other consequential action
Enforce entitlement at retrieval, before content reaches the model
Deploy via sandbox to production, not by editing production
Escalation and abandonment rates by topic are the useful production signal

Traps

Agents cannot return data the executing context lacks rights to. The platform does not silently bypass sharing rules.
Instructing a model to respect permissions is not a control.
Salesforce platform knowledge is the real prerequisite, not AI knowledge.
Data 360 is worth the same as prompt engineering — 20% each.

Night-before checklist

Six sections and weights; AI Agents is 35%
Passing score is 72%, a real target
Security model: entitlement enforced at retrieval
Check ID and proctoring rules — see exam day

Take the 20-question practice test.

AI-103 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AI-103. Everything here is covered in depth in the other guides in this section; this page is for the night before.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Price (US)	$165
Replaces	AI-102, retired 30 June 2026

Domain weights

Skill area	Weight
Implement generative AI and agentic solutions	30–35%
Plan and manage an Azure AI solution	25–30%
Implement computer vision solutions	10–15%
Implement text analysis solutions	10–15%
Implement information extraction solutions	10–15%

The top two are over half the exam. Spend your time accordingly.

AI-901 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AI-901. This is a fundamentals exam: it tests whether you can match a scenario to the right capability, not whether you can build one.

Exam facts


Level	Fundamentals
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Replaces	AI-900

Skill areas

Area	Weight
Implement AI solutions by using Microsoft Foundry	55–60%
Identify AI concepts and capabilities	40–45%

Only two areas, and the Foundry one is the bigger half.

AIF-C01 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for AIF-C01. A foundational exam: match the business scenario to the right concept.

Exam facts


Duration	90 minutes
Questions	65 (50 scored)
Passing score	700 / 1000
Price	100 USD

Domain weights

Domain	Weight
Applications of Foundation Models	28%
Fundamentals of GenAI	24%
Fundamentals of AI and ML	20%
Guidelines for Responsible AI	14%
Security, Compliance, and Governance	14%

Responsible AI plus governance is 28% combined — as much as the largest single domain, and the most under-studied part.

CY0-001 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for CY0-001. Securing AI systems is 40% of the exam, and the right answer removes an attack’s mechanism rather than merely detecting it.

Exam facts


Duration	60 minutes
Questions	Max 60
Passing score	600 / 900
Launched	February 2026
Price	Not published by CompTIA; varies by region

Domain weights

Domain	Weight
Securing AI systems	40%
AI-assisted security	24%
AI governance, risk, and compliance	19%
Basic AI concepts related to cybersecurity	17%

Attack names

Attack	What it is
Prompt injection	Instructions smuggled into input; indirect arrives via retrieved content
Jailbreaking	Phrasing that circumvents guardrails
Data poisoning	Corrupting training data to plant behaviour
Model extraction	Many crafted queries to replicate the model
Membership inference	Determining whether a record was in training data
Adversarial evasion	Crafted input that defeats detection at inference
Denial of wallet	Expensive requests flooding a paid endpoint

Attack to control

Attack	Control that removes the mechanism
Indirect prompt injection	Isolate retrieved content from the instruction channel
Model extraction	Rate limiting, quotas, restrict output richness
Data poisoning	Provenance checks and validation of contributed data
Excessive agent damage	Least-privilege tool scoping + confirmation on destructive actions
Secret disclosure	Keep secrets outside the model’s reach entirely
Denial of wallet	Rate limits, input size caps, per-tenant quotas
Weight theft in use	Confidential computing

Governance essentials

Inventory first — owner, purpose, risk classification. Nothing else works without it.
Assess before deployment and on a schedule, not once.
The deploying organisation is accountable, regardless of who built the model.

Traps

Filtering keywords is not the fix for prompt injection; isolation is.
Instructing the model not to reveal secrets is not a control.
Auto-closing alerts hides false negatives — sample and review closures.
Output-only filtering leaves malicious input reaching the model, which may drive harmful tool use before any output exists.
This is not an entry point into security. It assumes the fundamentals.

Night-before checklist

Name every attack in the table above unprompted
For each, the control that removes its mechanism
Four domains and weights
Check ID and proctoring rules — see exam day

Take the 20-question practice test.

Data Engineer Professional cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Data Engineer Professional exam. This one tests production judgement, so most answers concern what survives contact with reality.

Exam facts


Questions	60 scored (outline from 9 October 2026)
Duration	120 minutes
Fee	200 USD
Validity	2 years
Passing score	Not published by Databricks

Section weights

Section	Weight
Developing code for data processing	23%
Cost & performance optimisation	15%
Data ingestion & acquisition (incl. sharing, federation)	12%
Data manipulation	12%
Monitoring and alerting	10%
Debugging and deploying	10%
Security and compliance	8%
Data governance	5%
Data modelling	5%

Production patterns

Requirement	Answer
Exactly-once effect across restarts	Checkpointed offsets + idempotent or transactional sink
Append inside `foreachBatch` duplicated after restart	foreachBatch is at-least-once; use batch ID or MERGE
Streaming aggregation in append mode	Needs a watermark; emits once the window closes
Python UDF dominating runtime on 400M rows	Built-in expressions or a vectorised UDF
Heavy spill joining 4 TB to 30 MB	Check why the small side is not broadcast
Identical behaviour for backfill and stream	One transformation, parameterised by read/write mode
Logic must be testable outside a notebook	Versioned module with unit tests
Cost tripled, data grew 20%, heavy listing	Small file accumulation → compact
Frequent filters on two high-cardinality columns	Liquid Clustering, not partition
Managed table, no one wants to pick layout keys	`CLUSTER BY AUTO` with Predictive Optimization
Many small updates rewriting large files	Deletion vectors
Downstream needs only changed rows, incl. deletes	Change Data Feed
JSON with a shifting shape	`VARIANT` + `parse_json` / `variant_get`
Streaming job silently behind at peak	Input backlog / source lag
Alert only when actionable	Alert on final failure after retries

Security and governance

Erasure request on an append-only table → retained versions and downstream copies must be addressed within the retention policy; with deletion vectors, rewrite the files (REORG ... APPLY (PURGE)) and VACUUM. Deleting the current row is not enough.
Mask every column with one governed tag, across all tables → ABAC policy on the tag, not per-table masks.
SELECT granted, query still denied → missing USE CATALOG / USE SCHEMA on the parents.
Service principal with workspace admin → excessive privilege; scope to what the job needs.
Impact of a schema change → lineage in the governance catalog, not a code search.
Make tables findable → tags and comments on the securable.
Divergent metric definitions → one governed definition, such as a Unity Catalog metric view, both consumers use.

Ingestion

CDC feed → ordered merge including deletes and late or out-of-order events
CDC from SQL Server, MySQL or PostgreSQL → Lakeflow Connect
Partner needs live data, no workspace accounts → OpenSharing (formerly Delta Sharing); joint analysis without exposing raw data → Clean Rooms
Query an external database without copying → Lakehouse Federation
Rate-limited API with partial pages → backoff, resumable pagination, completeness validation

The published-score trap

No passing score is published. At professional level the margin is narrower, so calibrate on a weighted practice test.

Databricks Data Analyst cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Data Analyst Associate exam.

Exam facts


Questions	45 scored
Duration	90 minutes
Fee	200 USD
Language	English only
Passing score	Not published by Databricks

Section weights

Section	Weight
Executing queries using Databricks SQL and SQL Warehouses	20%
Creating Dashboards and Visualizations	16%
Analyzing Queries	15%
Genie spaces	12%
Databricks Data + AI Platform	11%
Managing Data	8%
Securing Data	8%
Importing Data	5%

Warehouse behaviour


Serverless	Starts in seconds; suits intermittent interactive work
Classic	Slower start; same SQL and results
Auto-stop	Shuts down during idle gaps — the main cost control
Query profile	Where time and data volume actually go; always look here first

Performance quick answers

Slow dashboard scanning a huge table → filter on the partitioned date column so file skipping applies
Large shuffle joining a 50-row dimension → broadcast the small table
Query slow after many small writes → compact small files
Same query, different results between analysts → different catalog, schema or table version
Need the result as of last Tuesday → Delta time travel by timestamp or version

Security and modelling

Region-restricted rows, one shared table → row filter keyed to the user’s group
Sensitive column hidden from most analysts → column mask
Reusable, always-current result without a copy → a view
Recurring CSV drops → incremental ingestion tracking processed files
Numbers landing as strings → explicit schema or cast on read

Concepts

Medallion — bronze raw, silver cleansed, gold business-level
Delta Lake — ACID transactions and schema enforcement over object storage
Genie space quality comes from table and column comments, instructions and example queries

The published-score trap

Databricks publishes no passing score. Any percentage quoted elsewhere is a guess. Use a weighted practice test as your only calibration.

Databricks Data Engineer Associate sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Data Engineer Associate exam.

Exam facts


Questions	45 scored
Duration	90 minutes
Fee	200 USD
Validity	2 years
Passing score	Not published by Databricks

Section weights

Section	Weight
Data Transformation and Modeling	22%
Data Ingestion and Loading	21%
Working with Lakeflow Jobs	16%
Governance and Security	15%
Implementing CI/CD	10%
Troubleshooting, Monitoring, Optimization	10%
Databricks Intelligence Platform	6%

Ingestion plus transformation is 43%.

Databricks GenAI Engineer cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Generative AI Engineer Associate exam.

Exam facts


Questions	45 scored
Duration	90 minutes
Fee	200 USD
Validity	2 years
Passing score	Not published by Databricks

Section weights

Section	Weight
Application Development	30%
Assembling and Deploying Apps	22%
Design Applications	14%
Data Preparation	14%
Evaluation and Monitoring	12%
Governance	8%

Development plus deployment is 52%.

RAG quick answers

Symptom or requirement	Answer
40,000 docs updated weekly	RAG over an index refreshed as docs change
Answers must cite the source paragraph	Retrieval returns identifiers and locations
Topically related but useless passages	Smaller chunks with overlap
Tables lose meaning when flattened	Preserve row/column relationships
Exact identifiers missed by semantic search	Hybrid search with keyword matching
Output must be fixed JSON	Constrain to a schema and validate
Latency dominated by 50 chunks, quality flat above 8	Retrieve fewer, rerank to keep the best
Hidden instructions in retrieved content	Isolate retrieved content from the instruction channel
Tool charges a customer	Explicit confirmation before execution

Deployment

Tenfold month-end spike → autoscaling with min and max
Compare two prompt versions on live traffic → split traffic, compare measured outcomes
Reproducible six months later → version the full configuration: prompt, model version, retrieval settings, code
Must not be publicly reachable → private network access, not an IP allow-list

Governance and evaluation

Entitlements enforced at retrieval, before content enters the context
Log metadata, source references and decisions, sensitive content redacted
Groundedness = supported by sources; relevance = addresses the question
Quality decline with no code change → drift in the question distribution or the corpus

The published-score trap

No passing score is published. Calibrate on a weighted practice test.

Databricks ML Associate cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Machine Learning Associate exam. Remember the balance: the platform section outweighs model development.

Exam facts


Questions	48 scored
Duration	90 minutes
Fee	200 USD
Validity	2 years
Passing score	Not published by Databricks

Section weights

Section	Weight
Databricks Machine Learning	38%
Model Development	31%
ML Workflows	19%
Model Deployment	12%

Platform concepts

Concept	What it is for
Experiment tracking	Parameters, metrics and artefacts per run, so results are comparable and reproducible
Model registry	Versioned stage transitions with an approval trail
Feature store	One feature definition shared by training and serving — the fix for skew
AutoML	A fast baseline and reviewable code, not a final unexamined answer
Environment spec	Library versions recorded with the run; the usual reason a run will not reproduce

Modelling quick answers

Random split on a time series → leaks the future; split chronologically
Feature computed from the target → target leakage; great validation, poor production
Encoder must match at serving → fit in training and persist
Rare positive class, 99.7% accuracy → precision, recall, AUC-PR
Train loss down, validation up → overfitting; regularise or stop early
Large errors disproportionately costly → RMSE
Wide variance across CV folds → unstable estimate, not a good model
Two models within 0.5% → take the simpler one
Why was this prediction made → per-prediction attribution, not global importance

Deployment

60M records overnight, no latency need → batch inference job
Accuracy declining over a year → data and concept drift

The published-score trap

No passing score is published. Any quoted percentage is a guess.

Databricks ML Professional cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Machine Learning Professional exam. Half of it is operations.

Exam facts


Questions	59 scored
Duration	120 minutes
Fee	200 USD
Language	English only
Passing score	Not published by Databricks

Section weights

Section	Weight
Model Development	44%
ML Ops	44%
Model Deployment	12%

MLOps answers

Requirement	Answer
Roll back within minutes	Registry retains the previous version + tested rollback
Labels arrive 30 days late	Proxy monitoring now, true metrics on arrival
Validate on live traffic without risking all users	Canary or shadow deployment
Automated retraining promoted a worse model	Promotion gate vs incumbent on a held-out set
Reconstruct a prediction six months later	Model version, served features and configuration for that request
Two teams break each other’s pipelines	Schema and semantics contract enforced in CI
Cost jumped after promotion	Serving instance type, min replicas, autoscaling bounds
Prevent unreviewed changes shipping	Code review and tests as mandatory gates

Development answers

Feature available at training but three days late in production → cannot serve as evaluated
Predictions diverge from validation → training-serving skew; one shared definition
Monthly rolling window retrain → parameterise the window, record the resolved data version
Strong aggregate, fails one segment → evaluate per segment
Two teams report different accuracy → different data, splits or metric definitions
Ensemble gains 0.3% at triple cost → decide explicitly against business value; default simpler
Reproducibility needs seeds, library versions, data version and non-deterministic ops together

Deployment

Sub-100ms with feature-store lookups → online feature lookup latency is the risk
Nightly 80M batch plus low-latency single requests → separate serving paths, same registered model version

Data and concept drift

Input distribution change versus a change in the input-to-target relationship. The second is only visible once labels arrive.

MLA-C02 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for MLA-C02. Remember the shape: three of four domains are operations, not modelling.

Exam facts


Passing score	720 / 1000
Level	Associate
Validity	3 years
Price	75 USD while in beta

Domain weights

Domain	Weight
Data Preparation for ML and AI	28%
ML Model and Foundation Model Development	24%
Deployment and Orchestration	24%
Operating, Monitoring, and Securing	24%

Failure modes

Symptom	Name	Fix
Great offline, poor in production	Target leakage	Remove features unavailable at inference
Predictions differ from validation	Training-serving skew	One shared feature definition
Accuracy declines over months	Data or concept drift	Monitor against the training baseline
99.7% accuracy, never predicts positive	Class imbalance	Precision, recall, AUC-PR; resample or weight
Wide variance across CV folds	Unstable estimate	Repeated or stratified CV

Requirement to answer

Requirement	Answer
5% of traffic, instant rollback	Canary with weighted routing
40M records overnight, no latency need	Batch inference job
Retrain when new labels land, with lineage	Event-triggered pipeline with artefact tracking
Traffic spikes at month end	Autoscaling with min and max
Training job reads one prefix	Execution role scoped to that prefix
Endpoint not on the public internet	VPC-private endpoint
Labels arrive two weeks late	Proxy monitoring now, true metrics later
Missing a positive case is costly	Recall
Large errors disproportionately costly	RMSE, not MAE

Traps

Random splits leak the future on time series. Split chronologically.
Fit encoders in training and persist them; re-fitting at serving changes the mapping.
Equal performance, simpler model wins. Marginal metric gains rarely justify operational complexity.
Interruptible capacity with checkpointing cuts cost without risking quality; halving the dataset does not.

Night-before checklist

Four domains, near-equal weight
Skew versus drift versus leakage, cold
Batch versus real-time serving decision
Check ID and proctoring rules — see exam day

Take the 20-question practice test.

N10-009 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for N10-009. Troubleshooting is the heaviest domain, and most troubleshooting questions hand you the clue that eliminates three options.

Exam facts


Duration	90 minutes
Questions	Max 90
Passing score	720 / 900
Version	V9, superseding N10-008
Price	Not published by CompTIA; varies by region

Domain weights

Domain	Weight
Network troubleshooting	24%
Networking concepts	23%
Network implementation	20%
Network operations	19%
Network security	14%

Subnetting quick table

CIDR	Block size	Usable hosts
/30	4	2
/29	8	6
/28	16	14
/27	32	30
/26	64	62
/25	128	126
/24	256	254
/23	512	510
/22	1024	1022

Network address = the block boundary below the host. 10.20.5.130/26 → blocks at .0 .64 .128 .192 → network is 10.20.5.128.

SC-500 cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for SC-500. The four areas carry near-equal weight, so a blind spot in any one is expensive.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Replaces	AZ-500

Skill areas

Area	Weight
Secure storage, databases, and networking	25–30%
Manage identity, access, and governance	20–25%
Secure compute	20–25%
Manage and monitor security posture	20–25%

Requirement to control

Requirement	Control
No credentials stored anywhere	Managed identity
Elevated access for a few hours a month	Time-bound, approval-gated activation
Extra verification only for risky sign-ins	Risk-based conditional access
Reachable only from one virtual network	Private endpoint + public access disabled
Partner read access for seven days, no account	Scoped, time-limited user delegation SAS
Revoke access to data instantly	Customer-managed keys
Interactive VM access, no public management ports	Managed bastion host
Only approved executables may run	Application control with an allow-list
Data protected from the host during processing	Confidential computing
Baseline enforced on every new subscription	Policy at management group scope
Correlate alerts across products	SIEM
Track configuration improvement over time	Secure score

Traps

A firewall rule allowing your office IP still exposes a public endpoint. Private endpoint plus disabled public access is the answer.
Disabling an account does not revoke live tokens. Sessions and tokens must be revoked explicitly.
Access reviews, not activity logs, answer “is this access still needed”.
Scan images in the pipeline and registry, not only running containers.
Tiered retention beats deleting logs at 30 days when investigations reach back months.
0.0.0.0/0 is never an acceptable debugging step. Grant a scoped, time-limited rule instead.

Night-before checklist

Four areas, roughly equal weight — name them
Private endpoint versus firewall rule, cold
Managed identity as the default credential answer
Check ID and proctoring rules — see exam day

Take the 20-question practice test and check your weakest quarter.

Spark Developer Associate cheat sheet

Sun, 20 Sep 2026 00:00:00 +0000

Last-minute reference for the Databricks Associate Developer for Apache Spark exam.

Exam facts


Questions	45 scored
Duration	90 minutes
Fee	200 USD
Language	Python
Passing score	Not published by Databricks

Section weights

Section	Weight
DataFrame / DataSet API applications	30%
Spark architecture and components	20%
Using Spark SQL	20%
Troubleshooting and tuning	10%
Structured Streaming	10%
Spark Connect	5%
Pandas API on Spark	5%

Transformations vs actions

Transformations are lazy — select, filter, withColumn, join. They build the plan. Actions trigger execution — count, collect, show, write.

Is 200-901 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

200-901 is moderately hard, and the difficulty depends on where you start. For a network engineer who has never programmed it is hard, because it asks you to read Python, JSON, Dockerfiles and Git output. For a developer with no networking background it is also hard, because a fifth of the exam is networking and Cisco platforms. Cisco publishes no pass rate or passing score, so no official difficulty figure exists; the difficulty comes from breadth and from reading code in a time limit.

Is AB-100 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

AB-100 is hard for people who have only built agents and have not designed or governed them, and moderate for experienced solution architects. It is an expert-level exam, and Microsoft requires a qualifying associate certification as well, so it sits above the builder exams. The difficulty comes from breadth across platforms, from scenarios where several answers work, and from the weight on operating and securing AI, not from deep configuration.

Is AB-730 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

AB-730 is one of the easier Microsoft exams on paper: it is a beginner-level exam with no coding and a 45-minute limit. It is not trivial, because it asks about exact Copilot features and about judgement, and because part of the outline is recent. If you use Microsoft 365 Copilot every day and read the study guide, expect a manageable exam; if you have only tried a chat assistant, expect to need real preparation.

Is AB-731 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

AB-731 is moderate for most business leaders: it needs no coding and the content is broad rather than deep, but the scenario wording and look-alike Microsoft products catch people who skim. Microsoft rates the level as Beginner. For someone who already leads change and has used Copilot, four weeks of part-time study is a reasonable plan; for someone new to AI, allow five or six. Microsoft publishes no pass rate, so any figure you see is a guess.

Is AZ-305 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

AZ-305 is hard in a particular way: it is an expert-level design exam where several answers are technically valid and only one fits the stated requirement. It is harder than AZ-104 for most people, not because the services are deeper, but because you must weigh trade-offs such as cost, recovery objectives and operational overhead. With solid AZ-104 knowledge and four weeks of focused study, it is very passable.

What makes it hard

Several answers are right

A typical question lists four real services that could all do the job in some scenario. The deciding detail is a requirement: read access to the secondary region, minimal administration, no code changes, the lowest cost. People who know every service but skim the requirement lose marks. People who read the requirement first usually do not.

Is AZ-400 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

AZ-400 is a hard expert exam for anyone who does not write pipelines regularly, and a moderate one for engineers who do. The difficulty comes from three things: more than half the exam is build and release pipelines, you must know two toolchains (GitHub and Azure DevOps), and many questions ask you to read YAML and judge it quickly. The content is not mathematically hard, but it is wide and hands-on.

Is CCNA 200-301 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

The CCNA is moderately hard: harder than entry-level exams such as Network+ because it expects you to configure and troubleshoot, but within reach of anyone who builds labs. Cisco does not publish a pass rate or passing score, so no official figure for difficulty exists. What makes it demanding is breadth, speed on subnetting, and reading command output under a 120-minute limit.

Why people find it hard

Factor	Why it hurts
Breadth	Addressing, switching, routing, services, security and now AI in one exam
Subnetting speed	Many topics assume you can subnet instantly
Output reading	v2.0 objectives are phrased as diagnose, troubleshoot and interpret
Details	Native VLANs, LACP modes, wildcard masks and administrative distance
Cost	A retake costs another exam fee (300 USD at time of writing)

Difficulty by domain

Domain	Weight	Difficulty	Main hurdle
Network Infrastructure and Connectivity	25%	Medium	Subnetting and IPv6 under time pressure
Switching and Network Access	25%	Medium to high	Rapid PVST+ roles and EtherChannel details
IP Routing	20%	Medium	OSPF adjacency and DR election
Network Services and Security	20%	Medium to high	ACL wildcard masks, NAT terms, Layer 2 security
AI and Network Operations	10%	Low to medium	New material; mostly definitions and severity levels

Spanning tree is the topic most candidates call the hardest, because it requires a diagram, not a rule. ACLs come second: one wrong wildcard bit or a missing permit changes the answer.

Is DP-600 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

DP-600 is a moderately hard, associate-level exam. It is harder than a fundamentals exam because it combines three languages (SQL, KQL and DAX) with design decisions, and it is easier than expert-level exams because the scope is one platform. Microsoft rates it intermediate. People who already build models in Power BI or write T-SQL for a living tend to find it manageable; people new to Fabric find the breadth the main obstacle.

Is DP-700 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

DP-700 is a moderately hard associate exam: harder than a fundamentals exam, code-heavy compared with analytics-focused exams, and much harder for anyone who has read about Fabric without using it. The difficulty is not one deep topic. It is breadth, three languages (SQL, PySpark and KQL), and scenarios in which several options are plausible.

What makes it hard

Three equal domains

Each domain is 30–35%, so you cannot rely on one strength. Implement and manage, ingest and transform, and monitor and optimize each ask for different habits: configuration and governance, coding, and diagnosis. Few engineers are strong at all three.

Is DP-750 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

DP-750 is a moderately hard associate exam: harder than a fundamentals exam, comfortable for engineers who run Databricks pipelines in production, and tough for people who have only used notebooks. Microsoft publishes no pass rate, so any figure you see quoted is a guess. The difficulty comes from breadth, from reading code, and from the production topics that learning on a notebook never forces you to meet.

What makes it hard

It is wide

Four domains cover compute, governance, six ingestion routes, modelling, transformation, quality, jobs, Git, bundles, Spark tuning and Azure Monitor. Few people use all of that in one job, so most candidates have blind spots.

Is DP-800 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

DP-800 is a moderately hard associate exam: harder than a fundamentals exam, and hardest for people who know either databases or AI but not both. Difficulty comes from breadth. In 120 minutes you need T-SQL fluency, database security, tuning, source-controlled deployment and an understanding of vector search and RAG. Microsoft does not publish a pass rate, so any figure you see for it is not an official number.

What makes it hard

It mixes two worlds

Developers who know T-SQL well often have not used the vector type, external models or embeddings. AI engineers know embeddings and RAG but may never have read an execution plan, written row-level security or built a database project. The exam expects both halves, and the database half carries more weight.

Is PL-300 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

PL-300 is a moderately hard associate exam: easier than the engineering-level exams, harder than a fundamentals test, and much harder for people who have built reports by dragging fields without learning how the model works. The difficulty does not come from the volume of features. It comes from DAX filter context, from modelling decisions that look similar, and from questions that describe a requirement and offer four plausible options.

What makes it hard

DAX is a different way of thinking

Many candidates come from Excel. DAX looks like Excel formulas, but a measure does not sit in a cell: it is evaluated once per cell of a visual, under a set of filters that come from rows, columns, slicers and relationships. CALCULATE changes those filters. Until you can predict what a measure returns for a given visual, scenario questions feel like guessing. This is the most common reason people fail the modelling domain.

Is SC-200 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

SC-200 is a moderately hard associate exam: harder than the fundamentals SC-900, comparable with other associate security exams, and much easier if you already work in a SOC. The difficulty comes from three things: two large products to know in configuration detail, KQL you must read and write, and scenario questions with several plausible answers. Nothing about it is mysterious, but reading alone rarely passes it.

What makes it hard

Factor	Why it bites
Breadth of products	Defender for Endpoint, Office 365, Identity, Cloud Apps, Entra, Purview and Sentinel all feature
KQL	You must write or complete queries, and know which table holds which signal
Configuration depth	Data collection rules, retention tiers, device groups and automation levels are rarely used by analysts day to day
Moving target	The English exam updates on October 21, 2026, and new topics such as data lake tiers and Sentinel Graph appear
Scenario wording	Several options work; one satisfies the stated constraint

What makes it easier

A free practice assessment from Microsoft that shows the style of the questions.
A published outline that lists every objective.
A lab you can build cheaply, so most concepts can be tried rather than memorised.
Overlap between domains: KQL practice helps all three.

Who finds it easier

Working SOC analysts know incident handling well. Their gaps are usually configuration objectives and the newest Sentinel features.

Is SC-300 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

SC-300 is a moderately hard associate exam. It is harder than a fundamentals exam such as SC-900, because it tests configuration decisions rather than definitions, and it is easier than an expert-level design exam. Candidates who work with Microsoft Entra ID every day usually find it fair; candidates who only know it from reading find the scenario questions slippery. Microsoft does not publish pass rates, so no official number tells you otherwise.

Is SC-401 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

SC-401 is a moderately hard associate exam: harder than a fundamentals exam such as SC-900, and much harder for people who have read about Microsoft Purview but never configured it. The difficulty does not come from depth in one product. It comes from many look-alike features, from prerequisites hidden in the scenario, and from three equally weighted domains that leave nowhere to hide.

What makes it hard

Look-alike features

The exam constantly asks you to pick between two things that each sound plausible: a retention policy or a retention label; a sensitivity label or a retention label; EDM or a trainable classifier; Exchange DLP or Endpoint DLP; an auto-labeling policy or a default label. Knowing both definitions is not enough. You must spot the requirement word that decides, such as per item, whole location, structured values or on a device.

Is SC-900 hard? Exam difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 is one of the easier Microsoft exams: it is a fundamentals exam with no configuration, no code and a 45-minute limit. It is still not trivial. The difficulty comes from breadth and similar names, not depth: dozens of Defender, Entra and Purview features that sound alike, and statement questions where one word makes a choice wrong. Most candidates with a few weeks of steady study pass; people who skim and rely on common sense are the ones who struggle.

Is SY0-801 hard? Difficulty explained

Fri, 09 Oct 2026 00:00:00 +0000

SY0-801 is moderately difficult. CompTIA positions Security+ as an entry-level certification for security administrators with two years of hands-on experience, and the main challenge is breadth, not depth. You need to recognise a large vocabulary of controls, attacks, protocols and documents, and pick the best answer from options that all sound plausible. People with IT experience usually pass after a few weeks of focused study. People with none often find it hard.

Free AZ-802 practice test: 20 questions

Tue, 29 Sep 2026 00:00:00 +0000

Twenty questions across the seven AZ-802 skill areas, weighted roughly as the real exam is: five on AD DS, two on hybrid management, two on virtual machines, two on networking, four on storage and file services, two on security and three on monitoring and troubleshooting. These are original scenario questions. Find the constraint in each scenario before you look at the options.

Deploy and manage AD DS

Question 1. New user accounts fail to be created on one domain controller with an error about the RID pool, while other DCs work. The DC holding a certain FSMO role was decommissioned without transferring its roles. Which role must you seize?

AI-300 cheat sheet

Mon, 28 Sep 2026 00:00:00 +0000

Last-minute reference for AI-300. The model lifecycle domain is the largest; the three generative AI domains together are worth 40–55%.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Language	English only
Replaces	DP-100

Skill areas

Area	Weight
Implement machine learning model lifecycle and operations	25–30%
Design and implement a GenAIOps infrastructure	20–25%
Design and implement an MLOps infrastructure	15–20%
Implement generative AI quality assurance and observability	10–15%
Optimize generative AI systems and model performance	10–15%

Requirement to answer

Requirement	Answer
Recreate a workspace identically per environment	Bicep + Azure CLI
No stored Azure secret in GitHub	OIDC federated credential
Approval before prod deployment	GitHub environment with required reviewers
Same model or environment in dev, test and prod	Registry
Read storage without keys	Identity-based datastore + managed identity
Training compute that costs nothing idle	Compute cluster, min nodes 0
Track parameters and metrics with one line	mlflow.autolog()
Search a continuous range and stop bad runs	Random sampling + bandit policy
Explain predictions, find failing cohorts	Responsible AI dashboard
Real-time scoring	Managed online endpoint
Score millions of files overnight	Batch endpoint
Test on real traffic, no user impact	Mirrored traffic
Canary release and instant rollback	Traffic split between deployments
Alert when inputs change	Model monitor with data drift signal
App calls a model without a key	Managed identity + RBAC on Foundry
Steady high volume, predictable latency and cost	Provisioned throughput (PTUs)
Data must stay in a geography	Regional or data zone deployment, not global
Keep the tested model version	Pinned version, no auto-upgrade
Answer contains facts not in the context	Groundedness
Answer doesn’t address the question	Relevance
Harmful content before launch	Risk and safety evaluators, adversarial data
Why did this one request fail?	Tracing in Application Insights
Exact codes missed by vector search	Hybrid search
Consistent style or format	Supervised fine-tuning
Preferred vs rejected answer pairs	DPO

Sweep sampling

Sampling	Continuous values	Early termination
Grid	No	Yes
Random	Yes	Yes
Bayesian	Yes	No

RAG levers

Chunk size and overlap, top-k, similarity threshold, hybrid search, semantic ranking, embedding model. Changing the embedding model means re-embedding everything.

Is AZ-104 hard? Exam difficulty explained

Mon, 28 Sep 2026 00:00:00 +0000

AZ-104 is a moderately hard associate exam: clearly harder than a fundamentals exam, easier than the expert-level AZ-305, and much harder for people who have read about Azure without using it. The difficulty does not come from any one service being deep. It comes from breadth — five domains and dozens of services in 100 minutes — from networking scenarios that combine several settings, and from commands and templates you have to read quickly.

SEE Part 1 cheat sheet

Mon, 21 Sep 2026 00:00:00 +0000

A last-minute reference for SEE Part 1 (Individuals). Everything is for tax year 2025, which is the year the July 2026 – February 2027 exams test. Figures are listed only where they are confirmed on irs.gov for 2025.

Exam facts


Questions	100: 85 scored, 15 unscored
Time	3.5 hours, three locked sections (1–34, 35–67, 68–100)
Score	Scaled 200–800, pass at 500
Fee	$317 per part
Tax law	As amended through December 31, 2025
Materials	Closed book, on-screen calculator

Scored questions per domain

Domain	Qs	Share
Preliminary Work and Taxpayer Data	14	~16%
Income and Assets	17	20%
Deductions and Credits	17	20%
Taxation	15	~18%
Advising the Individual Taxpayer	11	~13%
Specialized Returns for Individuals	11	~13%

Verified 2025 figures

Item	2025
Standard deduction: single / MFS	$15,750
Standard deduction: MFJ / QSS	$31,500
Standard deduction: HOH	$23,625
State and local tax limit	$40,000 ($20,000 MFS), reduced at higher incomes
Child tax credit	$2,200 per child, up to $1,700 refundable
IRA contribution	$7,000 ($8,000 age 50+)
Gift annual exclusion	$19,000 per recipient
Estate basic exclusion	$13,990,000
Home sale exclusion	$250,000 / $500,000 joint
FBAR	Foreign accounts over $10,000 aggregate
Form 8938, single in US	Over $50,000 year-end or $75,000 any time

Rules worth knowing cold

Marital status is set on December 31; a spouse who died during the year still allows a joint return for that year.
Qualifying child age: under 19, under 24 if a full-time student, any age if disabled.
Tiebreaker: parent over non-parent; then longer residence; then higher AGI.
Gifted property basis: donor’s basis for gain, FMV for loss if lower; a sale in between gives no gain or loss.
Inherited property: FMV at death, always long-term.
Wash sale: repurchase within 30 days, loss added to new basis.
Medical deduction: only above 7.5% of AGI.
Home equity interest: deductible only if used to buy, build or improve the home.
Child tax credit: child under 17 at year-end; otherwise the credit for other dependents.
Self-employment tax: on 92.35% of net earnings; half is an adjustment to income.
Estimated tax safe harbor: 90% of this year or 100% of last year (110% at higher incomes).
NIIT: 3.8% on the lesser of net investment income or MAGI above the threshold; IRA distributions excluded.
Housing allowance: excluded from income tax, included for SE tax.
Refund claim: 3 years from filing or 2 years from payment, whichever is later.
Injured spouse (Form 8379) is about a refund offset; innocent spouse (Form 8857) is about an understatement.
Form 706: due 9 months after death; portability requires a Form 706 even if no tax is due.
FBAR goes to FinCEN, not with the return; Form 8938 goes with the return.

Traps

Old or new figures. A number from 2024 or 2026 is wrong on this exam.
EXCEPT questions. Mark each option true or false first.
Filing status hidden in the facts. Settle it before calculating anything.
Locked sections. Review flagged questions before leaving each section.

Night-before checklist

PTIN and appointment confirmed; ID ready
Remote: system check done, desk clear, no paper allowed
The verified figures table, once through
Check ID and proctoring rules — see exam day

Take the 20-question practice test.

Is 200-901 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

200-901 is worth it if you want a recognised credential that you can combine network knowledge with code: you automate networks, write against Cisco APIs, or move from network operations towards software and DevOps practice. It is less worth it if your job is purely hands-on switching and routing, where the CCNA fits better, or purely application development with no Cisco platforms involved.

What the certification proves

Cisco describes the exam as testing software development and design, understanding and using APIs, application deployment and security, and infrastructure and automation on Cisco platforms. Cisco states no prerequisites, and the certification is valid for three years. Cisco does not publish salary figures in the pages used for this guide, so this guide gives none.

Is AB-100 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

AB-100 is worth it if your job is to decide how AI agents are used in business solutions on Microsoft platforms: architects, senior consultants and technical leads who plan, design and govern them. It is less worth it if you mostly build agents or write application code, where an associate-level exam fits your role better. Microsoft maps the certification to the job role Solution Architect.

Who it is for

Microsoft’s audience profile describes accomplished solution architects with expertise in designing and delivering AI-driven business solutions. The ideal audience listed for its related course includes solution and enterprise architects, senior Dynamics 365, Microsoft 365, Power Platform and Azure AI consultants, AI and digital transformation leads, and application architects and technical leads.

Is AB-730 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

AB-730 is worth taking if you use Microsoft 365 daily and want an independent, recognised way to show that you use Copilot and agents effectively and carefully. It is less useful as a route into a technical career, because it deliberately tests no coding, administration or architecture. Microsoft publishes no salary or job-outcome data for it, so judge it on the skills it proves and the effort it takes.

What the certification shows

Microsoft describes the credential as for people who use generative AI productivity tools and core Microsoft 365 apps to improve business outcomes and decision-making. In practice it signals that you can:

Is AB-731 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

AB-731 is worth taking if you lead or influence AI adoption and want a credential that shows you understand the business case, the Microsoft tools and the governance, without needing to code. Microsoft positions the Microsoft Certified: AI Transformation Leader credential for business decision-makers at all levels. If your role is technical, or you simply use Copilot as a user, another exam may serve you better. This page uses only what Microsoft publishes; it makes no salary claims.

Is AZ-305 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

AZ-305 is worth it if you design, advise on or lead Azure solutions, or if you already hold AZ-104 and want the certification that represents the architect role. It leads to the Azure Solutions Architect Expert credential, which Microsoft describes as the role of advising stakeholders and translating business requirements into Azure designs. It is less worth it if you mainly build pipelines or write application code, where other exams fit better.

Is AZ-400 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

AZ-400 is worth it if you build and operate delivery pipelines on Azure or GitHub, or you want a recognised credential for that work. It is Microsoft’s expert certification for the DevOps engineer role, and it covers the work of that role: source control, pipelines, deployments, infrastructure as code, security scanning and monitoring. It is less worth it if your work is mostly application code with no involvement in delivery, or pure infrastructure design.

Is DP-600 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

DP-600 is worth it if you build or manage analytics on Microsoft Fabric: lakehouses, warehouses and Power BI semantic models. It is Microsoft’s role-based certification for the Fabric analytics engineer, and it proves you can prepare data, secure and govern it, and build semantic models at enterprise scale. It is less worth it if your work is only building report visuals, or only ingesting data with Spark, where PL-300 or DP-700 fit better.

Is DP-700 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

DP-700 is worth it if you build or run data engineering solutions on Microsoft Fabric, or your organisation is adopting Fabric and you want your skills evidenced. It is Microsoft’s associate-level certification for the Fabric data engineer role and covers ingestion, transformation, orchestration, security and monitoring. It is less worth it if you work in a platform other than Fabric, or if your job is building reports rather than pipelines.

Is DP-750 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

DP-750 is worth taking if you build or run data pipelines on Azure Databricks and want a Microsoft credential that says so. It is not worth taking if your platform is Microsoft Fabric, if you are new to data engineering, or if you only need Databricks credentials from the vendor. The deciding question is whether Azure Databricks is where your work happens.

Who it is for

Microsoft describes the candidate as a data engineer with expertise in integrating and modelling data, building and deploying optimised pipelines, and troubleshooting and maintaining Azure Databricks workloads, with experience in data quality and governance in Unity Catalog. It fits:

Is DP-800 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

DP-800 is worth it if you build or maintain applications on SQL Server, Azure SQL or SQL databases in Microsoft Fabric and want to show that you can add AI features to them securely. It validates a specific skill set: T-SQL and database design, security, tuning, CI/CD and vector search with RAG. It is less worth it if your work is mostly Python-based AI applications or large-scale lakehouse pipelines. This page uses only facts Microsoft publishes; it makes no salary or hiring claims.

Is PL-300 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

PL-300 is worth it if you build, or want to build, reports and semantic models in Power BI for other people. It is Microsoft’s associate certification for the Power BI data analyst role, it covers the whole loop from raw data to a secured, published report, and it is a recognised way to show that you can do the work. It is less worth it if your job is mainly SQL, pipelines or platform engineering, where DP-600 or DP-700 fit better.

Is SC-200 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

SC-200 is worth taking if your job or target job is in security operations: triaging alerts, investigating incidents, hunting for threats or engineering detections with Microsoft tools. It is a role-based associate credential, and Microsoft describes the role it validates clearly. It is less useful if you work mainly on building preventive controls or on compliance, where other exams fit better.

What the certification shows

Passing earns the Microsoft Certified: Security Operations Analyst Associate credential. According to Microsoft, the holder is a security operations analyst who reduces organisational risk by triaging, responding to incidents, hunting for threats and engineering detections, using Defender XDR, Sentinel, Entra ID, Purview and Defender for Cloud workload protections, and performing hunting with KQL.

Is SC-300 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

SC-300 is worth it if you administer or plan to administer identity in a Microsoft environment: Entra ID, hybrid sync, Conditional Access, application access and privileged access. It is Microsoft’s associate certification for that role and covers the work in the role description. It is less worth it if your daily work is data protection, security operations or application code, where other exams fit better.

What the credential says

Microsoft describes the identity and access administrator as someone who designs, implements and operates an organisation’s identity and access management with Microsoft Entra, covering identities for users, devices, Azure resources and applications, and applying Zero Trust principles. The certification page lists the role as Security Engineer and the level as intermediate. It demonstrates that you can modernise identity solutions, implement hybrid identity and implement identity governance.

Is SC-401 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

SC-401 is worth it if your job involves protecting data in Microsoft 365: labelling it, preventing leaks, retaining it, investigating risky behaviour or answering legal discovery requests. It is Microsoft’s associate certification for the Information Security Administrator role and it follows the work of that role closely. It is less worth it if your work is incident response or identity administration, where SC-200 and SC-300 fit better.

Who it is for

Microsoft 365 and security administrators. Microsoft’s audience profile expects familiarity with Microsoft 365 services, PowerShell, Microsoft Entra, the Defender portal and Defender for Cloud Apps. If you already administer the tenant, SC-401 adds the data-protection layer.

Is SC-900 worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 is worth it if you work near security, identity or compliance on Microsoft platforms and want a recognised starting point, or if you plan to take one of the Microsoft security role-based exams and want the vocabulary first. It is less worth it if you are already preparing for an associate-level security exam and have the knowledge, because fundamentals certifications are not required for those exams.

Who it is for

Microsoft’s audience profile names business stakeholders, new or existing IT professionals and students, all with an interest in Microsoft security, compliance and identity solutions. The certification page lists the level as beginner and the related role as security engineer, which describes where the path leads, not what SC-900 qualifies you to do.

Is SY0-801 Security+ worth it? Who should take it

Fri, 09 Oct 2026 00:00:00 +0000

Security+ SY0-801 is worth it if you are starting out in cybersecurity or in a systems or network role that includes security duties, and you want a vendor-neutral credential that proves a broad baseline. It is less worth it if you already have years of security experience and need a specialist credential, or if you only need depth in one vendor’s cloud. CompTIA aims it at security administrators with two years of hands-on experience, and its page names roles such as security analyst, security engineer and systems administrator.

Is the CCNA 200-301 worth it?

Fri, 09 Oct 2026 00:00:00 +0000

The CCNA is worth it if you want to work with enterprise networks, particularly Cisco-based ones, and want an associate credential that proves hands-on skill with switching, routing, security basics and now automation and AI. It is less worth it if your work never touches networking or if a vendor-neutral entry credential would do. The exam has no prerequisites, costs 300 USD at the time of writing, and the certification stays valid for three years.

AZ-802 study resources

Tue, 29 Sep 2026 00:00:00 +0000

The best AZ-802 resources are free: Microsoft’s study guide, the Windows Server and Azure documentation for each objective, evaluation copies of Windows Server for a lab, and an Azure subscription for the hybrid parts. Add realistic practice questions to find your weak domain. The one thing to be careful with is the large body of AZ-800 and AZ-801 material: much of it is still accurate, but a good share covers topics AZ-802 dropped.

200-901 career: from DevNet to NetDevOps

Fri, 09 Oct 2026 00:00:00 +0000

CCNA Automation (200-901) points towards network automation engineer, software developer on Cisco platforms and system integration programmer roles. Cisco’s own certification page names exactly those three as the audience, describing it as built for “entry-to-mid-level IT professionals” who want to specialise in software development and design. The credential is the old DevNet Associate under a new name, so job adverts, résumés and LinkedIn profiles may still say either “DevNet Associate” or “CCNA Automation” for the same thing. This page covers the roles, the reworked professional tier above it, and what a hiring manager can and cannot read into the badge.

AAIA career: auditor roles and next steps

Fri, 09 Oct 2026 00:00:00 +0000

AAIA does not open a new profession. It marks an existing IT auditor or advisor as someone who can give an opinion on AI systems. ISACA’s AAIA page names no job titles. It describes the audience as experienced IT auditors and advisors, people who assess, implement, maintain or audit AI systems, and holders of an active prerequisite designation. The prerequisite is the real career signal: you cannot hold AAIA without first holding CISA or another qualifying designation, so the credential always arrives as a second line on a CV, never the first.

AAIR career: AI risk roles and next steps

Fri, 09 Oct 2026 00:00:00 +0000

AAIR points at risk officers and IT risk professionals who now have AI on their register, not at a new job title. ISACA describes the audience as experienced IT risk professionals who work across functions on AI risk response, and the same page requires an active qualifying designation such as CISA, CISM, CRISC, CGEIT or CDPSE before the credential can be awarded. It lists no job titles of its own, so the roles below are an interpretation, labelled as such, rather than something ISACA promises. Details are on ISACA’s AAIR page.

AAISM career: CISO path and next steps

Fri, 09 Oct 2026 00:00:00 +0000

AAISM leads to the same jobs your CISM or CISSP already pointed at, with an AI remit written into them. ISACA’s own AAISM page names no job titles at all, and it describes the credential as something that supplements certified security managers. Its candidate guide describes the target as IT security professionals who hold a CISM or CISSP and want recognition for securing enterprise AI solutions and for using AI to improve security operations. So the question is less “what job does it unlock” and more “what does it add to the job I have”.

AB-100 career: architect role and next steps

Fri, 09 Oct 2026 00:00:00 +0000

AB-100 leads to one job role in Microsoft’s catalogue: Solution Architect. The certification page tags Microsoft Certified: Agentic AI Business Solutions Architect Expert with that single role, and the exam alone does not earn it, because Microsoft also demands one associate-level certification from a list of 13 on the certification page. This page looks at who that profile fits, what the credential does not prove, and which exams people usually add around it. For renewal mechanics see renewal and retakes.

AB-620 career: agent builder roles

Fri, 09 Oct 2026 00:00:00 +0000

Passing AB-620 earns the Microsoft Certified: AI Agent Builder Associate credential, and Microsoft tags it with two roles on its certification page: App Maker and Developer. The audience profile is more specific than either tag. It describes a “professional developer or advanced builder” who typically works as an IT application developer, a consultant or an independent software vendor (ISV) partner. No job title called “AI Agent Builder” is promised by the credential; it describes a body of work, not a position.

AB-730 career: what it does for business roles

Fri, 09 Oct 2026 00:00:00 +0000

AB-730 leads to no job title of its own: Microsoft tags the Microsoft Certified: AI Business Professional credential with the roles Business User and Administrator, at Beginner level, for the product Microsoft 365 Copilot. In practice it is a proof of working method for someone whose job is not IT: it shows that a person can use Copilot, Copilot Cowork and prebuilt agents carefully, without writing code. Whether that matters for a promotion or a new job depends on the employer, and Microsoft publishes no job-outcome data for it. The facts below come from the certification page and the study guide, both read on 2026-10-09.

AB-731 career: who it suits, what it leads to

Fri, 09 Oct 2026 00:00:00 +0000

AB-731 leads to the Microsoft Certified: AI Transformation Leader credential, which Microsoft tags with the role “Business Leader” and the level “Beginner”. It is a credential for people who decide and sponsor, not for people who build: Microsoft says the holder is “not expected to write any code”. It does not open a technical career path, and Microsoft’s pages name no follow-on exam. What it can do is document that a manager, programme lead or consultant knows the Microsoft AI toolset well enough to lead conversations about it. This page covers the roles it fits and the limits of that claim; validity is on the renewal and retakes page.

AB-900 career: Copilot admin roles

Fri, 09 Oct 2026 00:00:00 +0000

AB-900 leads to entry-level work on a Microsoft 365 admin team, not to a job title of its own. Microsoft files the credential under the Administrator role at Beginner level on Microsoft Learn, describes it as showing you can “support, secure, and protect an AI-enabled Microsoft 365 environment”, and lists no prerequisite exam; the credential does not expire. What it opens is a conversation with an employer, not a ladder rung. The renewal and retake rules are covered separately.

Agentforce Specialist career: roles and next steps

Fri, 09 Oct 2026 00:00:00 +0000

The Salesforce Certified Agentforce Specialist is a role-extension credential, not an entry ticket: Salesforce aims it at administrators, developers and architects who already work on the platform, and expects about a year of configuration experience including Data 360. It did not start life under this name. It was the AI Specialist credential until March 2025, when Salesforce renamed it and shifted the outline towards Agentforce. If you want to know what it leads to, the honest answer is that it adds an agent-building line to a Salesforce job you already have.

AI-103 career: Azure AI developer roles

Fri, 09 Oct 2026 00:00:00 +0000

AI-103 points at two job families. On its certification page Microsoft tags the Azure AI Apps and Agents Developer Associate credential with the roles AI Engineer and Developer, at Intermediate level, under the products Azure and Microsoft Foundry. The overview describes the candidate as “an Azure AI engineer who builds, manages, and deploys agents and AI solutions that take advantage of Microsoft Foundry”. The credential does not lead to a single fixed job; it supports applications for roles in which Python developers wire generative AI into products. The sources are Microsoft’s certification page and its AI-102 retirement page.

AI-200 career: developer roles it points to

Fri, 09 Oct 2026 00:00:00 +0000

AI-200 leads to the same job titles AZ-204 did, back-end and cloud developer roles on Azure, with an AI slant. Microsoft files the Azure AI Cloud Developer Associate under the role tag “Developer”, with the subjects “Application development” and “Artificial intelligence”, and its audience profile describes someone responsible for AI solutions “with an emphasis on back-end services and components”. It does not describe a data scientist or a prompt engineer. Details are on Microsoft’s certification page.

AI-300 career: from data scientist to MLOps

Fri, 09 Oct 2026 00:00:00 +0000

AI-300 leads to operations-side AI work on Azure: Microsoft’s own page tags the Machine Learning Operations Engineer Associate credential with the AI Engineer role, not the Data Scientist role that its predecessor DP-100 carried. The credential describes someone who sets up infrastructure for machine learning operations (MLOps) and generative AI operations (GenAIOps), which Microsoft groups together as AI operations, or AIOps. A pass is a statement about running models in production, and it is read best next to a portfolio of pipelines you can explain. The renewal and retake rules matter here too, because the credential lapses after a year unless it is renewed.

AI-901 career: what the credential leads to

Fri, 09 Oct 2026 00:00:00 +0000

AI-901 leads to an entry point rather than a job title. Microsoft’s certification page for Microsoft Certified: Azure AI Fundamentals lists exactly one job role, AI Engineer, and one required exam, AI-901. The audience line is worded for someone “at the beginning of your career in AI solution development”, which is a narrower and more technical description than the one the retired AI-900 carried. The credential is a first marker for that path, not a qualification for the role itself.

AIF-C01 career: who it suits and what next

Fri, 09 Oct 2026 00:00:00 +0000

AIF-C01 leads to no single job title. AWS lists it for six broad groups (cloud, development, data, IT, AI/ML and line-of-business roles) and describes the ideal candidate as someone who “uses, but does not necessarily build” AI/ML solutions on AWS. It proves you can choose and explain AI approaches, not that you can ship them. The details are on AWS’s AI Practitioner certification page.

Six audiences, one common thread

AWS names the audience by function rather than by job title. In practice that covers a wide spread of real roles:

AZ-104 career: roles and next steps

Fri, 09 Oct 2026 00:00:00 +0000

AZ-104 leads to Azure administrator work: the person who keeps an organization’s Azure subscriptions, networks, storage, virtual machines and identities running. Microsoft files the Azure Administrator Associate credential under the single role tag “Administrator” at the “Intermediate” level, and it is the only entry-level-to-expert bridge on the Azure infrastructure track, because the Azure Solutions Architect Expert credential cannot be earned without it. This page covers the jobs the badge points at, the limits of what it shows, and the exams that usually follow. Whether to take it at all is covered in Is AZ-104 worth it?.

AZ-305 career: architect roles, next steps

Fri, 09 Oct 2026 00:00:00 +0000

AZ-305 leads to solution architect work on Azure: the person who advises stakeholders and turns business requirements into designs. Microsoft files the Azure Solutions Architect Expert credential under one role tag, “Solution Architect”, at expert level, and the credential is made of two parts: the Azure Administrator Associate certification plus a pass on AZ-305. This page covers the jobs that wording points at, what the expert badge does and does not show, and what usually follows. For the decision to sit the exam at all, see Is AZ-305 worth it?.

AZ-400 career: DevOps engineer roles

Fri, 09 Oct 2026 00:00:00 +0000

AZ-400 leads to one job family on Microsoft’s own map: the DevOps engineer. Microsoft files the DevOps Engineer Expert certification under the single job role “DevOps Engineer”, describes that person as a developer or infrastructure administrator, and says the certification has “2 certifications” as prerequisites, which means AZ-400 alone is never enough. This page covers what the title means in practice, who the role works with, what the credential does not show, and which exam tends to come next. Whether the exam is worth the effort is answered in Is AZ-400 worth it?.

AZ-802 career: Windows Server admin roles

Fri, 09 Oct 2026 00:00:00 +0000

AZ-802 leads to Windows Server administration work: the person who runs domain controllers, Hyper-V hosts, file servers and the Azure Arc, Update Manager and Monitor tooling around them. Microsoft’s certification page tags the Windows Server Administrator Associate credential with seven roles at once (Administrator, Identity and Access Administrator, Information Protection and Compliance Administrator, Network Engineer, Security Engineer, Support Engineer and Technology Manager), which is more than most associate certifications carry. This page reads what that list does and does not mean for a job search. Whether the exam is worth the effort is covered in Is AZ-802 worth it?, and the validity rules are in renewal and retake rules.

AZ-900 career: who it helps and what follows

Fri, 09 Oct 2026 00:00:00 +0000

AZ-900 leads to very little on its own, and that is by design: Microsoft lists the Azure Fundamentals certification at Beginner level and calls it “a common starting point in a journey towards a career in Azure”. The certification page tags the role as Administrator, yet the audience profile is broader than that: any technology professional who works in “infrastructure management, database management or software development” and wants a shared vocabulary for Azure. This page covers who the credential suits, what it cannot show, and where people usually go afterwards.

CCNA 200-301 career: jobs and next steps

Fri, 09 Oct 2026 00:00:00 +0000

The CCNA 200-301 leads to entry-level and mid-level networking work: network technician, network administrator and junior network engineer. Cisco’s own certification page describes the CCNA as built for “entry-to-mid-level IT professionals” who want to specialise in networking, security or automation, and names aspiring network engineers, administrators and technicians as the target group. It is a single-exam, associate-level credential with no prerequisites, so it is a starting rung rather than a senior qualification. The practical value depends on what you add next: real device time, a specialty, or the CCNP.

CY0-001 career: where SecAI+ fits in a job

Fri, 09 Oct 2026 00:00:00 +0000

CY0-001 does not lead to a job title of its own. CompTIA says so in plain words in its blog post on jobs you could get with SecAI+: the credential “does not promise a job title, and it does not replace role-based security credentials.” What it does is add an AI layer to jobs that already exist (SOC analyst, cloud security engineer, DevSecOps engineer, GRC analyst), and it is aimed at mid-level practitioners rather than newcomers. The exam launched on February 17, 2026, as the first certification in what CompTIA calls its expansion series. Whether it is worth the money is covered in Is CY0-001 worth it?; this page covers the roles, the limits and the exams that sit around it.

Databricks Data Analyst Associate careers

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Data Analyst Associate points at one job family: people who query, chart and explain data inside Databricks SQL. Databricks files it under a “Data Analyst” role on its certification page but publishes no audience profile or list of job titles, so every title below is interpretation. What the page does state is narrow and useful: no prerequisites, 6+ months of hands-on analysis experience recommended, 45 multiple-choice questions, and a recertification requirement every two years.

Databricks Data Engineer Associate career

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Data Engineer Associate supports a move into data engineer roles on the Databricks platform, but Databricks itself names no job titles for it. Its certification page describes the exam as assessing foundational data engineering skills, lists no prerequisites (“related training highly recommended”), and ends with a two-year expiry. Everything below about roles is therefore interpretation, and it is labelled as such.

What Databricks says, and what it leaves out

The vendor wording is short. The exam assesses “introductory data engineering tasks” on the Databricks Intelligence Platform, and Databricks recommends hands-on experience with the tasks in the exam guide. No audience profile, no list of job roles and no seniority claim is published on that page.

Databricks Data Engineer Pro career paths

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Data Engineer Professional is pitched at people who build, optimise and maintain production pipelines, but Databricks names no job titles for it. Its certification page says the exam validates “advanced skills in building, optimising, and maintaining production-grade data engineering solutions”, and that a passing candidate “can be expected to complete advanced data engineering tasks”. Any mapping to titles such as senior or lead data engineer is interpretation, and this page labels it that way.

Databricks GenAI Engineer career paths

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Generative AI Engineer Associate points at one kind of work: building and deploying RAG applications and LLM chains on Databricks. Databricks’ certification page says people who pass “can be expected to build and deploy performant RAG applications and LLM chains”, but it publishes no job titles, so every title below is interpretation. The credential is an Associate with no prerequisites, and it expires after two years.

How Databricks words the target role

The exam page describes an individual who can “design and implement LLM-enabled solutions using Databricks”. Two phrases carry the meaning. The first is problem decomposition: breaking complex requirements into manageable tasks. The second is choosing “appropriate models, tools and approaches from the current generative AI landscape”. Neither describes research. Both describe an engineer assembling existing components.

Databricks ML Associate: jobs it leads to

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Machine Learning Associate sits on the first rung of Databricks’ “ML Engineer” track, yet its own wording is about using the platform for “basic machine learning tasks”, not about a job title. On the certification page the credential appears under an ML Engineer heading with Associate and Professional tiers, while the exam text promises that a passer can complete basic ML tasks with AutoML, Unity Catalog and select features of MLflow. Databricks publishes no list of job roles, so every title below is interpretation.

Databricks ML Professional career: ML engineer

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Machine Learning Professional is filed on Databricks’ certification landing page under the ML Engineer role, at the Professional level, and the exam page says passers can be expected to perform “advanced machine learning engineering tasks at enterprise scale”. That makes the realistic target a senior or mid-to-senior machine learning engineer or MLOps engineer on Databricks, not a data scientist whose work ends in a notebook. Both statements come from the Databricks exam page and the certification overview.

DP-600 career: the analytics engineer role

Fri, 09 Oct 2026 00:00:00 +0000

DP-600 leads to the Microsoft Certified: Fabric Analytics Engineer Associate credential, which Microsoft’s own catalogue files under two role tags at once: Data Engineer and Data Analyst. That double tag is the best summary of the job. The analytics engineer is neither pure pipeline builder nor pure report author, and the credential suits people who already work in the gap between the two. The certification page lists the level as Intermediate and the product as Microsoft Fabric. Details of the exam itself are in the exam format page; renewal rules are in DP-600 renewal and retakes.

DP-700 career: Fabric data engineer roles

Fri, 09 Oct 2026 00:00:00 +0000

DP-700 leads to one credential, Microsoft Certified: Fabric Data Engineer Associate, and Microsoft tags it with exactly one role, “Data Engineer”, and one product, Microsoft Fabric. The certification page describes a person with subject matter expertise in data loading patterns, data architectures and orchestration processes. It is an intermediate-level credential, and the English exam content is revised on October 19, 2026.

The role as Microsoft words it

On the certification page the Fabric data engineer has three responsibilities: ingesting and transforming data, securing and managing an analytics solution, and monitoring and optimizing it. The same page says this person works closely with analytics engineers, architects, analysts and administrators to design and deploy data engineering solutions for analytics, and should be skilled at manipulating data with SQL, PySpark and KQL.

DP-750 career: Azure Databricks roles

Fri, 09 Oct 2026 00:00:00 +0000

DP-750 leads to the Microsoft Certified: Azure Databricks Data Engineer Associate credential, and Microsoft files it under exactly one role tag, Data Engineer, with the product tag Azure Databricks and the level Intermediate. The credential is not a ticket to a new job title. It is evidence for people who already build or run pipelines on the Azure flavour of Databricks. The details below come from the certification page and the study guide, and the fee, formats and renewal are covered on renewal and retakes.

DP-800 career: SQL developers adding AI

Fri, 09 Oct 2026 00:00:00 +0000

DP-800 leads to developer work that sits on a SQL database rather than beside it. Microsoft tags the Microsoft Certified: SQL AI Developer Associate credential with the role “Developer”, the level “Intermediate”, the product “Azure” and the subject “Data management” on its certification page. There is no separate “database engineer” or “AI engineer” tag, so the credential reads as a SQL developer who can also ship vector search and retrieval-augmented generation (RAG). Whether the exam is worth the effort is covered in Is DP-800 worth it?.

MLA-C02 career: ML engineer roles on AWS

Fri, 09 Oct 2026 00:00:00 +0000

MLA-C02 maps to a short list of titles that AWS itself publishes: ML engineer, MLOps engineer, LLMOps engineer, data engineer, backend software developer and data scientist. The “LLMOps engineer” tag is new compared with the previous version of the exam, and it signals what AWS now expects from the role: running foundation models in production, not only classic models. The wording comes from AWS’s certification page, and the full task list is in the MLA-C02 exam guide.

N10-009 career: help desk to network tech

Fri, 09 Oct 2026 00:00:00 +0000

Network+ N10-009 is aimed at the step between help desk and infrastructure work. CompTIA’s certification page says it prepares candidates for jobs in technical support, network operation and system administration, and CompTIA’s own N10-009 blog names junior network administrator, NOC technician, network support specialist and systems administrator as the primary roles. The exam assumes A+ plus 9 to 12 months of hands-on experience, so it works as a ladder rung for people already inside IT. Dates, fees and expiry are on the renewal and retake page.

PL-300 career: BI analyst roles and next steps

Fri, 09 Oct 2026 00:00:00 +0000

PL-300 leads to report-building and semantic-model work in Power BI: BI analyst, Power BI developer, reporting analyst and similar titles. Microsoft files the Power BI Data Analyst Associate credential under one role tag, “Data Analyst”, at the “Intermediate” level, and its audience profile is written around delivering insight to business stakeholders, not around running a data platform. The credential is renewed annually, which is covered in PL-300 renewal and retake rules. This page covers the jobs it points at, what it leaves unproven, and where people usually go next.

SC-200 career: SOC roles and next steps

Fri, 09 Oct 2026 00:00:00 +0000

SC-200 leads to security operations work built on Microsoft tooling: alert triage, incident response, threat hunting and detection engineering. Microsoft’s own role label for the credential is “Security Operations Analyst”, and its description says the holder triages, responds to incidents, hunts for threats and engineers detections with Defender XDR, Sentinel, Entra ID, Purview and Defender for Cloud. That sentence is the best guide to the jobs the exam supports, and it covers more than the entry-level desk.

SC-300 career: identity admin roles

Fri, 09 Oct 2026 00:00:00 +0000

SC-300 leads to identity and access administration work in Microsoft environments, and Microsoft tags the credential with a single role: Security Engineer, at the Intermediate level, under the subject “Security”. That tag is worth noticing, because the exam title says “administrator” while the catalogue filing says “engineer”. The certification page describes someone who designs, implements and operates identity and access management with Microsoft Entra, and the English version of that description is updated together with the exam on October 28, 2026.

SC-401 career: data protection roles

Fri, 09 Oct 2026 00:00:00 +0000

SC-401 points to the person who configures data protection inside a Microsoft 365 tenant: sensitivity labels, data loss prevention, retention, insider risk and audit. Microsoft tags the Information Security Administrator Associate credential with one role, “Administrator”, at the “Intermediate” level, and files it under the subject “Information protection and governance” rather than under general security. That filing is the main clue to the jobs: this is a compliance-engineering credential, not a security operations one. It also follows an older, now retired Microsoft 365 compliance certification, which affects how a CV reads. The upkeep rules are on the renewal and retake page.

SC-500 career: security engineer roles

Fri, 09 Oct 2026 00:00:00 +0000

SC-500 leads to cloud security engineering: implementing the controls that protect identities, data, applications, infrastructure and AI workloads across Azure, Microsoft 365 and hybrid estates. Microsoft tags the Cloud and AI Security Engineer Associate credential with exactly one role, “Security Engineer”, at the “Intermediate” level, and lists “Generative AI” among its subjects, which the retired Azure Security Engineer Associate page did not. That retired credential was withdrawn on August 31, 2026, so this is now the only Microsoft associate credential for the role. Whether to take it is covered in Is SC-500 worth it?; this page covers where it points.

SC-900 career: what it leads to for newcomers

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 will not get you hired as a security engineer, and Microsoft does not claim it will. The certification page tags it at the Beginner level with a single related role, Security Engineer, and describes the audience as business stakeholders, new or existing IT professionals and students. That makes it a credential that opens a conversation, not one that fills a job title. Whether it helps depends on which of those three people you are.

SEE Part 1 career: tax preparer to EA

Fri, 09 Oct 2026 00:00:00 +0000

Passing SEE Part 1 does not make you an enrolled agent and does not by itself change what you may do for clients. It is one of three exam hurdles on the way to a credential the IRS itself issues, and the IRS says an enrolled agent, like an attorney or CPA, has unlimited practice rights: any taxpayer, any tax matter, any IRS office. Until you have passed Parts 2 and 3, applied on Form 23 and cleared the suitability check, you are a Part 1 passer, and that is a study milestone rather than a job title.

SEE Part 2 career: business tax roles

Fri, 09 Oct 2026 00:00:00 +0000

A passed SEE Part 2 shows tested knowledge of business taxation, but it carries no job title and no new rights: the IRS grants enrolled agent status only after all three parts are passed and Form 23 is approved. What Part 2 does is mark you as someone who can handle sole proprietors, partnerships, C and S corporations, trusts and estates, exempt organizations and retirement plans, which is the content of the three Part 2 domains (30, 37 and 18 scored questions). The IRS describes the finished credential as “the highest credential the IRS awards”, with unlimited practice rights before any IRS office.

SEE Part 3 career: representation work

Fri, 09 Oct 2026 00:00:00 +0000

Passing SEE Part 3 is the point where the exam stops being about computing tax and starts being about the right to stand in for a client. It still does not make you an enrolled agent on its own: the IRS says candidates apply on Form 23 within one year of the third pass and then clear a suitability check covering tax compliance and a criminal background check. Only after that does the IRS describe the holder as someone with “unlimited practice rights”, the same as attorneys and CPAs, meaning any taxpayer, any tax matter and any IRS office (see the IRS enrolled agent page).

Spark Developer Associate: jobs it supports

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Associate Developer for Apache Spark supports roles where you write PySpark, but Databricks names no job titles for it. The certification page describes the target as someone who can complete “basic Spark DataFrame tasks” in Python, and the word “basic” is the most useful clue about seniority. Anything below about titles is interpretation, and it is labelled as such.

What the vendor wording does and does not say

Databricks assesses two things: understanding of Spark architecture and components, and the ability to apply the DataFrame API to basic data manipulation inside a Spark session. The page gives no audience profile, no role list and no seniority claim. It lists no prerequisites, recommends 6+ months of hands-on work, and sets a two-year validity.

SY0-801 career: security admin roles

Fri, 09 Oct 2026 00:00:00 +0000

Security+ SY0-801 leads mainly to hands-on junior and mid-level security administration work. CompTIA’s certification page lists three job roles for Security+, the same for V7 and V8: Security Analyst, Security Engineer and Systems Administrator. The V8 page recommends two years of hands-on experience as a security administrator, so the credential is designed to confirm skills people already use on the job, not to replace that experience. For dates, renewal and retakes, see SY0-801 renewal and retake rules.

AZ-802 cheat sheet

Tue, 29 Sep 2026 00:00:00 +0000

Last-minute reference for AZ-802. AD DS is the largest area; storage has the most objectives; troubleshooting ties everything together.

Exam facts


Duration	120 minutes
Passing score	700 / 1000
Price	Set per country or region; shown at booking
Languages	English
Renewal	Free, every 12 months

Skill areas

Area	Weight
Deploy and manage AD DS	20–25%
Manage storage and file services	15–20%
Monitor and troubleshoot Windows Server environments	15–20%
Manage Windows Server instances and workloads in a hybrid environment	10–15%
Manage virtual machines	10–15%
Implement and manage an on-premises and hybrid networking infrastructure	10–15%
Secure Windows Server infrastructure	10–15%

Requirement to answer

Requirement	Answer
Branch DC that stores only some passwords	RODC + Password Replication Policy
New objects fail, RID pool exhausted	RID master
Time source for the forest	PDC emulator of the forest root
Access to only some servers in another forest	Forest trust + selective authentication
One service account on several servers	gMSA (needs KDS root key)
User settings based on the computer	Loopback processing
Remote command reaches a second server	Resource-based Kerberos constrained delegation
Users may run only certain cmdlets	JEA
Manage non-Azure servers from Azure	Azure Arc
Scheduled patching across Azure and Arc	Azure Update Manager
Hyper-V inside a VM	Nested virtualization
Whole GPU for one VM / shared GPU	DDA / GPU partitioning
Test a replica without interruption	Hyper-V Replica test failover
Time-limited RDP to Azure VMs	Just-in-time VM access
On-premises resolves Azure private zone	Private Resolver inbound endpoint + conditional forwarder
Different answers per client subnet	DNS policy with zone scope
Backup DHCP server that waits	DHCP failover, hot standby
Small local disk, large file share	Azure File Sync + cloud tiering
File shares over the internet without VPN	SMB over QUIC
Block file types on a share	FSRM active file screen
Zero-loss volume replication	Storage Replica, synchronous
Unique local admin passwords	Windows LAPS
Baseline that fixes its own drift	OSConfig
Different password rules for a group	Fine-grained password policy
No NTLM, no caching for admins	Protected Users
Restore deleted user with memberships	AD Recycle Bin
Trust relationship failed	Test-ComputerSecureChannel -Repair
Forecast disk capacity locally	System Insights

Numbers worth knowing


Kerberos clock skew	5 minutes
Default intersite replication interval	180 minutes; minimum 15
Hyper-V Replica frequencies	30 seconds, 5 minutes, 15 minutes
Protected Users TGT lifetime	4 hours
SMB over QUIC	UDP 443, TLS 1.3
Azure platform DNS resolver	168.63.129.16

Traps

Seized FSMO holders must never return to the network.
Denied beats Allowed in the RODC Password Replication Policy.
Planned failover needs the primary VM off; test failover does not.
Deduplication cannot run on system or boot volumes; ReFS cannot be a boot volume.
The Recycle Bin cannot be disabled and does not cover objects deleted before it was enabled.
Only the forest root PDC emulator should sync with an external time source.
Credential Guard is not a replacement for Protected Users; one protects the machine, the other the account.
DHCP servers must be authorised in AD before they lease.

Night-before checklist

Name the seven areas and which three are largest
The five FSMO roles, their scope and symptoms, cold
Authoritative versus non-authoritative, for both AD and SYSVOL
Check ID and proctoring rules; see exam day

Take the 20-question practice test and check your weakest area.

200-901 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

CCNA Automation, the certification earned by passing 200-901, is valid for three years. Cisco’s CCNA Automation page lists two ways to renew within that period: pass the current 200-901 exam or any higher-level Cisco certification exam, or earn 30 continuing education (CE) credits. After a failed attempt the wait is five calendar days; after a pass you must wait at least 180 days before retaking the same exam number. This page applies those rules to the automation track and to the holders who earned the credential under its former name, DevNet Associate.

AAIA renewal, CPE and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

AAIA has no expiry date and no renewal exam. ISACA keeps the credential alive through annual CPE reporting and an annual fee: at least 10 CPE hours a year, at least 30 over a three-year reporting period, and US$20 for members or US$35 for non-members, due by 1 January. The CPE hours must be in the AI specialised domain. If you fail the exam rather than lapse after it, the limit is four attempts in a rolling twelve months. Figures come from ISACA’s AAIA maintenance page and the AAIA exam candidate guide, a 2025 edition.

AAIR renewal, CPE and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

AAIR never expires on a fixed date; it stays valid while three things continue: at least 10 CPE hours reported each year (30 over a three-year reporting period), the annual maintenance fee, and an active qualifying designation. ISACA’s AAIR maintenance page sets the fee at US$45 for members and US$85 for non-members, due by 1 January. Fail the requirements and the certification is revoked, which is different from a quiet lapse.

AAISM renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

AAISM is maintained, not retaken. ISACA asks holders for a minimum of 10 CPE hours a year and 30 over a three-year period, plus an annual fee of US$20 for members or US$35 for non-members, due by 1 January. Those numbers are a third of the 20 and 120 hours that CISM carries, which is the first thing that surprises people who assume an advanced credential is heavier to keep. If you fail the exam instead, ISACA allows four attempts in a rolling 12 months, with 30 and 90 day waits. All figures below come from ISACA’s AAISM maintenance page and the AAISM exam candidate guide (version 1.26).

AB-100 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Agentic AI Business Solutions Architect Expert credential is valid for one year and renews free through an online assessment on Microsoft Learn, but the key AB-100 quirk is that its associate prerequisite does not renew along with it. Microsoft’s FAQ says an expert certification does not automatically expire when the associate prerequisite expires, and that each certification must be renewed separately. If you fail the exam itself, you wait 24 hours before the second attempt and 14 days between later ones. Everything below is taken from Microsoft Learn pages checked on October 9, 2026. See also what AB-100 leads to.

AB-620 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The AI Agent Builder Associate earned with AB-620 is valid for one year. Microsoft’s credential expiration policy counts that year from the date all requirements are completed, and renewal is a free, unproctored online assessment on Microsoft Learn that opens six months before the expiry date. If you fail the exam itself, you wait 24 hours before the second attempt and 14 days between later ones.

Timeline from pass to expiry

Point in time	What Microsoft states
Day 0, exam passed	The certification is valid for one year
About month 6	The renewal assessment becomes available; reminder emails go out if your Learn profile has an email address and you agreed to communications
Months 6 to 12	Take the assessment any time; passing extends the certification one year from the expiration date
After expiry	You cannot renew; the certification must be earned again by passing the exam

Passing early does not shorten anything: the extension is counted from the existing expiration date, not from the day you passed. Expiry dates and times are in UTC, which matters if you leave the assessment to the last evening. Microsoft’s renewal page and its FAQ both recommend finishing well before the deadline.

AB-730 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft’s AB-730 certification page does not state whether Microsoft Certified: AI Business Professional expires, and that is the open question on this page. The retake rules, by contrast, are fixed: wait 24 hours after a first failed attempt, then 14 days between later attempts, with a maximum of five attempts in 12 months, paying each time. The English exam also changes on October 20, 2026. Everything below was read on Microsoft Learn on 2026-10-09; the sources are the certification page, the retake policy and the renewal FAQ.

AB-731 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft does not state on the AB-731 pages whether the AI Transformation Leader credential expires. What it does publish is a rule by level: associate, expert and specialty certifications are valid for one year, and fundamentals certifications do not expire. AB-731 is labelled “Beginner”, which is neither of those words, so the expiry term is not confirmed by an official AB-731 statement. The retake rules are clear: 24 hours after a first failure, then 14 days between later attempts, with a maximum of five attempts in 12 months. For career context see AB-731 career.

AB-900 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

AB-900 does not expire and has no renewal assessment, but a failed attempt follows a fixed schedule: 24 hours before the second try, then 14 days between all later tries, with at most five attempts in 12 months. Microsoft states both rules on its own pages: its renewal FAQ says fundamentals certifications “do not expire”, and its retake policy sets the waiting periods. The career page covers where the credential leads.

Why there is nothing to renew

Microsoft’s expiration policy gives each family its own rule. Associate, expert and specialty certifications last one year and can be renewed free of charge in the six months before expiry through an unproctored, open-book assessment on Microsoft Learn. Fundamentals certifications, which include Microsoft 365 Certified: Copilot and Agent Administration Fundamentals, are in a separate bucket that does not expire. The AB-900 study guide repeats the associate rule in its “useful links” table, which is the only renewal text in the exam’s own study materials.

Agentforce Specialist renewal and retakes

Fri, 09 Oct 2026 00:00:00 +0000

The Salesforce Certified Agentforce Specialist does not renew by retaking the exam. You keep it by completing one Trailhead maintenance badge per year, and Salesforce’s schedule currently puts the Agentforce Specialist badge in the Summer ‘26 cycle: available from August 2026, due August 2027. Miss the due date and the credential expires. There is no fee for the badge itself, as far as Salesforce’s published pages show.

How the yearly badge works

Salesforce’s Certification Maintenance Schedule states three rules that matter here: each certification needs one maintenance badge a year, badges appear with the Winter, Spring or Summer release and fall due a year later, and you must complete every module in the badge. After you finish, allow up to 24 hours for your certification status to update.

AI-103 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Azure AI Apps and Agents Developer Associate credential from AI-103 is valid for one year. Microsoft renews it free of charge through an online, unproctored, open-book assessment on Microsoft Learn, which becomes available six months before the expiry date. If it expires, it cannot be renewed and AI-103 has to be passed again. The failed-exam rule is separate: a 24-hour wait after the first fail, then 14 days between later attempts, with at most five attempts in 12 months. Sources: Microsoft’s credential expiration policy, renewal FAQ and retake policy.

AI-200 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Azure AI Cloud Developer Associate earned with AI-200 is valid for one year from the day all requirements are met, and it renews for free through an online assessment that opens six months before expiry. A failed AI-200 attempt can be retaken after 24 hours, then every 14 days, with at most five attempts in 12 months. These are Microsoft’s general rules for role-based certifications, not AI-200 specials, as set out on its renewal page and its retake policy.

AI-300 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The AI-300 certification is valid for one year, and it renews for free through an unproctored, open-book online assessment on Microsoft Learn that opens six months before expiry. If you fail the exam itself, you wait 24 hours for the second attempt, 14 days for each later one, and you may sit it at most five times in 12 months. These rules come from Microsoft’s general credential policies, which apply to AI-300 as to other role-based exams. Microsoft does not publish AI-300-specific renewal terms, and no separate fee or schedule applies to this exam. The career side is covered in what AI-300 leads to.

AI-901 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Azure AI Fundamentals does not expire, so AI-901 has no renewal assessment and no renewal deadline. What does have rules is failing: Microsoft requires a 24-hour wait after a first failed attempt, then 14 days between each later attempt, with no more than five attempts in 12 months. Those numbers come from Microsoft’s retake policy page as of October 2026, and a retake is paid again where a fee applies.

AIF-C01 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

AWS Certified AI Practitioner is valid for three years. You renew it by passing the latest AIF-C01 exam, the Machine Learning Engineer Associate exam or the Generative AI Developer Professional exam. If you fail, you wait 14 calendar days and pay the full fee again, with no limit on attempts. Both rules come from AWS’s recertification page and after-testing policy.

The three-year clock and what restarts it

AWS’s recertification page lists exactly three renewal routes for AI Practitioner, each giving three more years:

AZ-104 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Azure Administrator Associate credential you earn by passing AZ-104 is valid for one year from the day you complete all requirements, and you renew it for free by passing an unproctored online assessment on Microsoft Learn during the six months before it expires. If you fail the exam itself, you can retake it after 24 hours, then after 14 days between later attempts, with a cap of five tries in twelve months. Miss the renewal deadline and the credential cannot be renewed any more: you sit AZ-104 again. Careers context is on the career page; the exam mechanics are in exam format.

AZ-305 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Azure Solutions Architect Expert credential earned with AZ-305 is valid for one year and renews for free, through an unproctored online assessment on Microsoft Learn that opens six months before expiry. Retaking AZ-305 itself after a failed attempt follows a different rule: 24 hours after the first failure, then 14 days between later attempts, with a cap of five attempts in 12 months. The two systems are separate, and mixing them up is the common mistake. For jobs and next exams, see the career page.

AZ-400 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The DevOps Engineer Expert credential earned with AZ-400 is valid for one year and renews for free through an unproctored online assessment on Microsoft Learn that Microsoft says takes about 45 minutes. Two things set it apart from a plain associate renewal. It sits on top of a prerequisite credential that has its own separate expiry, and one of its two named prerequisites, Azure Developer Associate, is now retired. A failed AZ-400 exam can be retaken after 24 hours, with 14 days between later attempts. The career side is on the career page.

AZ-802 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Passing AZ-802 earns the Microsoft Certified: Windows Server Administrator Associate certification, valid for one year from the day you complete all requirements. You renew it for free with an unproctored online assessment on Microsoft Learn, available during the six months before expiry. If you fail the exam, you can retake after 24 hours, then after 14 days between later attempts, up to five tries in 12 months. This page adds the parts that are specific to Windows Server: the retirement of AZ-800 and AZ-801 on 30 September 2026, a renewal page that still carries the old name, and a rename that confuses holders. Career context is on the career page, and the exam mechanics are in exam format.

AZ-900 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Certified: Azure Fundamentals does not expire, so there is nothing to renew. The rules that do apply to AZ-900 are the retake limits: wait 24 hours after a first failed attempt, then 14 days between later attempts, with at most five attempts in 12 months. Microsoft publishes both rules on its Learn site, and this page walks through them as of 2026-10-09. The mechanism is the same for every Microsoft fundamentals exam; the details about the AZ-900 outline are specific to this one.

CCNA 200-301 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

A CCNA is valid for three years, and you can renew it before it expires either by passing the current CCNA exam or any higher-level Cisco certification exam, or by earning 30 continuing education (CE) credits. If you fail the 200-301 exam you must wait five calendar days before the next attempt, and after a pass you cannot take the same exam number again for 180 days. These rules come from Cisco’s recertification policy page and its CCNA page, and they are shared with most of Cisco’s associate, specialist and professional exams. The numbers that differ by level are the CE credits needed.

CY0-001 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

CompTIA SecAI+ is valid for three years and, according to CompTIA’s own tables, renews with 15 continuing education units (CEUs) and a $75 fee for the whole cycle. That is a lighter load than Security+, which needs 50 CEUs and a $150 fee in the same tables. For retakes there is no waiting period before the second attempt, a 14-calendar-day wait before the third and later ones, and the full exam price every time. The exam itself also has a built-in end date, since CompTIA estimates retirement about three years after the February 17, 2026 launch. Career context is on the career page; format and scoring are in exam format.

Databricks Data Analyst Associate renewal

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Data Analyst Associate stays valid for two years from the award date, and renewing means sitting the full current live exam again at the standard $200 fee. After a fail there is no attempt cap, but 14 days must pass between attempts and every attempt is paid. The sources are the exam’s certification page and the Databricks certification FAQ.

The two-year clock for the analyst badge

The exam page says recertification is required every two years to maintain certified status. The FAQ adds that the badge expires two years after it was awarded and that Databricks notifies holders before a certification expires. The clock runs from the award date, not from the day you booked.

Databricks Data Engineer Associate renewal

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Data Engineer Associate is valid for two years from the date it is awarded, and the only way to renew it is to take the full current live exam again. If you fail, you may retake as often as you like, but you must wait 14 days between attempts and pay the fee each time. These rules come from the Databricks certification FAQ.

The two-year clock

Databricks words it plainly: a certification is valid for two years from the date it is awarded, and all digital badges associated with its exams expire two years after they were awarded. The date that matters is the award date shown on your badge, not the exam date you booked.

Databricks Data Engineer Pro renewal rules

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Data Engineer Professional is valid for two years, and the only renewal route is sitting “the current version of the exam” again at the $200 registration fee. After a failed attempt you wait 14 days, and no attempt is free. The exam page states the validity and recertification lines, and the certification FAQ supplies the retake and rescheduling rules.

Rules that apply to this exam, line by line

Question	What Databricks publishes	Source
Validity	2 years from the award date	FAQ and exam page
Recertification	Required every two years; take the full current live exam	Exam page and FAQ
Fee	$200, the same figure the FAQ gives for all Databricks certification exams	FAQ
Notice before expiry	You will be notified in advance	FAQ
Retake wait	14 days between each take	FAQ
Attempt cap	None; retake “as many times as you would like”	FAQ
Retake pricing	Same as the original; no free retake vouchers	FAQ
Passing score	Not published, because it changes as questions are updated	FAQ

Everything in that table is identical for the other Databricks exams, including the Associate. This exam differs only in its content, which is the part that decides what a recertification attempt looks like.

Databricks GenAI Engineer renewal and retakes

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Generative AI Engineer Associate is valid for two years from the date it is awarded, and the only renewal route is to take the current version of the exam again. A failed attempt can be retaken after 14 days, any number of times, at the original $200 fee each time. The exam page and the certification FAQ state these rules, and they are the same for every Databricks exam.

Databricks ML Associate renewal rules

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Machine Learning Associate is valid for two years from the award date. To recertify you take “the current version of the exam”, and if you fail you may retake after 14 days, paying the full $200 each time. The rules come from the exam page and the certification FAQ, and they match the other Databricks exams, so this page concentrates on what is particular to the ML Associate.

Reading the two-year clock for this exam

The FAQ says a certification is valid for two years from the date it is awarded and that the digital badge expires two years after it was awarded. It also says holders are notified in advance of expiry. Databricks publishes no renewal window, no grace period and no early-renewal discount, so a lapsed holder simply sits the exam again.

Databricks ML Professional renewal and retakes

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Machine Learning Professional is valid for two years, and the only renewal route is to pass the current version of the exam again at the full $200 fee. After a failed attempt there is no attempt limit, but Databricks requires a 14-day wait and a new payment each time. The validity and recertification wording is on the exam page, and the retake wording is in the certification FAQ.

A two-year timeline for a Professional holder

When	What happens	Source
Day 0	Pass; unofficial result appears at the end, then Databricks reviews the proctor recording	FAQ
Day 0 to a few days	Badge posted at credentials.databricks.com, emailed through Accredible	FAQ
Before month 24	Databricks says it notifies you in advance of expiry	FAQ
Month 24	Credential and badge expire	FAQ
After month 24	Register again and sit the current exam, same fee	Exam page

The clock runs from the date the certification is awarded, not from the day you sat the paper. Databricks does not publish a renewal window, so whether sitting the exam a few months early extends the date to two years from the new award is not confirmed. Ask Databricks if a contract depends on it.

DP-600 renewal and retake rules (2026)

Fri, 09 Oct 2026 00:00:00 +0000

The Fabric Analytics Engineer Associate earned through DP-600 is valid for 12 months and is renewed for free by passing an unproctored online assessment on Microsoft Learn, not by retaking the exam. If you fail DP-600 itself you wait 24 hours before the second attempt and 14 days between later ones. The English exam and its skills outline change on October 19, 2026, and Microsoft states that renewal assessments follow exam updates. This page is based on Microsoft Learn pages loaded on 2026-10-09; see also DP-600 career paths.

DP-700 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Fabric Data Engineer Associate earned with DP-700 expires after one year, and Microsoft’s own page states that role-based and specialty certifications expire unless renewed, at no cost, by passing an online assessment on Microsoft Learn. The renewal window opens six months before expiry. A failed DP-700 attempt can be repeated after 24 hours, then after 14 days for later attempts. Sources: the certification page, the renewal page and the retake policy.

Timeline of a DP-700 credential

Moment	What Microsoft says
Pass DP-700	The Fabric Data Engineer Associate is awarded; it is an associate credential, not a fundamentals one.
Six months before expiry	The renewal assessment becomes available; it cannot be taken earlier.
Within the window	Pass the free, unproctored, open-book assessment; the certification is extended one year from the expiration date.
Expiry with no pass	The certification must be earned again by passing the required exam, here DP-700.

“Extended from the expiration date” means that renewing early does not shorten the year you receive. The FAQ also says the new expiration date shows on your Learn profile immediately after a pass. Microsoft recommends not leaving it to the last week, to allow for technical problems or several attempts.

DP-750 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Azure Databricks Data Engineer Associate earned with DP-750 is valid for one year, and you keep it by passing a free, unproctored, open-book renewal assessment on Microsoft Learn, which opens six months before the expiry date. If you fail the exam itself, you wait 24 hours, then 14 days between later attempts, with a maximum of five in 12 months. Databricks’s own credentials work the opposite way, as the Databricks Data Engineer Associate renewal page shows. Microsoft’s sources are the renewal page, its FAQ and the retake policy.

DP-800 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Microsoft Certified: SQL AI Developer Associate credential expires one year after you pass DP-800, and the only way to extend it is a free, unproctored renewal assessment on Microsoft Learn taken inside the six months before expiry. Failing the exam itself means a 24-hour wait, then 14 days between later attempts, with at most five attempts in 12 months. These are Microsoft’s general rules for associate exams, taken from its renewal FAQ and retake policy; nothing about DP-800 is treated differently. This page covers what is specific to the credential’s short history. Career context is on career and next steps.

MLA-C02 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

AWS Certified Machine Learning Engineer - Associate is valid for three years, and AWS lists three ways to extend it: pass the latest version of the exam (3 more years), pass the Generative AI Developer - Professional exam (3 more years), or “maintain” it on AWS Skill Builder (1 more year, only within 90 days of expiry). After a failed attempt you wait 14 calendar days and pay the full fee again. The rules come from AWS’s recertification page and after-testing policy.

N10-009 renewal, CEUs and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Network+ expires three years after you earn it, and CompTIA’s continuing education (CE) program asks for 30 continuing education units (CEUs) plus a $150 fee over that cycle. The 30-CEU requirement is lower than Security+ at 50 and higher than A+ at 20. The exam N10-009 launched on June 20, 2024, and CompTIA’s page gives an estimated retirement around 2027, so a holder who certified near launch will see the expiry date and the version change land close together. The roles side is in what N10-009 leads to.

PL-300 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Power BI Data Analyst Associate is valid for 12 months from the day you meet its requirements, and you keep it by passing a free, unproctored renewal assessment on Microsoft Learn before it expires. PL-300 failures follow a separate rule: wait 24 hours after the first failed attempt, then 14 days between later attempts, with at most five attempts in 12 months. Both sets of rules come from Microsoft’s own policy pages, summarised below with what they mean for a Power BI analyst. For the roles the credential points to, see PL-300 career.

SC-200 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Security Operations Analyst Associate earned with SC-200 is valid for 12 months and is renewed through a free, unproctored assessment on Microsoft Learn, not by sitting SC-200 again. Microsoft’s certification page lists the renewal frequency as 12 months. If you fail the exam itself, you can retake it after 24 hours. The English exam changes on October 21, 2026, which matters for anyone whose renewal window or study plan straddles that date.

SC-300 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Identity and Access Administrator Associate credential earned with SC-300 lasts 12 months, and renewal is free: pass an unproctored online assessment on Microsoft Learn before the expiry date. A failed SC-300 exam can be retaken after 24 hours, then after 14-day waits, up to five attempts in 12 months. Both rules come from Microsoft’s pages as read on October 9, 2026, and the exam itself changes on October 28, 2026, which matters for anyone planning a date near that line. The career side is in SC-300 career roles.

SC-401 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

SC-401 earns an associate credential that is valid for one year from the day all requirements are met, and it renews for free through an open-book assessment on Microsoft Learn. A failed exam can be retaken after 24 hours, then every 14 days, with at most five attempts in 12 months. The unusual part in autumn 2026 is that two Microsoft pages give different dates for the exam update, October 14 and October 28, which matters for both renewal and booking. For what the credential leads to, see the career page.

SC-500 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

The Cloud and AI Security Engineer Associate credential earned with SC-500 is valid for one year from the day all requirements are met, and renewal is a free, unproctored assessment on Microsoft Learn. If you fail the exam itself, you wait 24 hours before the second attempt and 14 days between later ones, with at most five attempts in 12 months. Those rules come from Microsoft’s policy pages, not from SC-500 specifically; the SC-500 details are in how they interact with a brand-new exam and a retired predecessor.

SC-900 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 does not expire, so no renewal is ever due: Microsoft states that fundamentals certifications do not expire, and the free annual renewal assessment applies only to associate, expert and specialty certifications. The rules that do matter for SC-900 are the retake limits (24 hours after a first failure, 14 days between later attempts, at most five attempts in 12 months) and what the October 21, 2026 exam update means for anyone mid-study. Everything below comes from Microsoft Learn pages checked on 2026-10-09.

SEE Part 1 retakes and pass validity

Fri, 09 Oct 2026 00:00:00 +0000

A passed SEE Part 1 does not expire in the way a vendor certificate does; it is a carryover credit that stays usable for three years from the date you passed, and the clock matters only until you have passed Parts 2 and 3. If you fail, you may retake Part 1 after 24 hours, up to four times per testing window, paying the full fee each time. All of this comes from the IRS’s enrolled agent FAQ, read on 2026-10-09.

SEE Part 2 expiry dates and retake limits

Fri, 09 Oct 2026 00:00:00 +0000

A passed SEE Part 2 does not renew: it is a credit that stays usable for three years from the day you pass, and it only matters until Parts 1 and 3 are passed too. The IRS’s enrolled agent FAQ (read 2026-10-09) allows each part four attempts per testing window, with a 24-hour wait after a fail before booking the same part again. The fee is $317 per sitting every time.

SEE Part 3 after passing: Form 23 and CE

Fri, 09 Oct 2026 00:00:00 +0000

After the third SEE pass, the rules change from retakes to enrollment: the IRS says to apply within one year of the date you passed the third part, using Form 23 (online through Pay.gov or on paper with a check), and to pass a suitability check. The enrollment fee is $140 according to the IRS FAQ, and the IRS asks you to allow 60 days for processing. Once enrolled, the credential must be renewed about every three years with 72 hours of continuing education. Facts here come from IRS pages read on 2026-10-09: the enrolled agent FAQ, the become an enrolled agent page and the maintenance page.

Spark Developer Associate renewal and retakes

Fri, 09 Oct 2026 00:00:00 +0000

The Databricks Certified Associate Developer for Apache Spark is valid for two years, and the only renewal route is to pass the full current exam again at the $200 registration fee. A failed attempt can be retaken without a cap, but only after 14 days and at full price. The rules come from the Databricks certification FAQ and the exam page, both read on 9 October 2026.

Two years, counted from the award date

The FAQ says a certification is valid for two years from the date it is awarded, and that digital badges expire two years after they were awarded. The exam page repeats it: “Recertification is required every two years.” The clock starts at the award date on the badge, not on the day you booked.

SY0-801 renewal and retake rules

Fri, 09 Oct 2026 00:00:00 +0000

Security+ does not last forever. CompTIA’s continuing education (CE) program asks for 50 continuing education units (CEUs) and a $150 fee over each three-year renewal cycle, and CompTIA’s retake policy lets you retry once without waiting but requires 14 calendar days before a third attempt. The new version SY0-801 is expected on or around November 17, 2026, while SY0-701 retires on June 11, 2027 (English). CompTIA’s pages do not publish a separate CE scheme for V8, so this page describes the program as it is written for Security+ now. The career side is in what SY0-801 leads to.

200-301 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for the CCNA 200-301 depends on how much of Cisco’s five-domain blueprint you already touch at work. The exam lasts 120 minutes and is built around subnetting, switching, routing, services and security, so a candidate who configures networks daily and a candidate who has never seen a switch need very different calendars, even though both sit the same paper. This guide uses the v2.0 weights: 25% infrastructure and connectivity, 25% switching and network access, 20% IP routing, 20% network services and security, and 10% AI and operations.

200-901 study scenarios: three starting points

Fri, 09 Oct 2026 00:00:00 +0000

The right way to study for 200-901 depends on which side of the network-and-software divide you come from. The exam lasts 120 minutes, is graded pass or fail, and weights six domains almost evenly: APIs (20%) and infrastructure and automation (20%) lead, while software development, Cisco platforms, deployment and security, and network fundamentals are 15% each. Because no domain is below 15%, nobody can skip a side, and each starting point leaves a different hole to fill.

AAIA study scenarios: three audit profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AAIA depends on where your audit experience stops and AI begins. The exam has 90 questions, and ISACA’s candidate guide gives 150 minutes and a pass mark of 450 on a scaled 200 to 800 range. Its weights are lopsided: AI Operations is 46%, AI Governance and Risk is 33%, and AI Auditing Tools and Techniques is only 21%. Because every candidate must already hold CISA or a qualifying audit designation held in an IT audit or advisory role, the open question is never audit method; it is which AI topics you are missing.

AAIR study scenarios: three risk profiles

Fri, 09 Oct 2026 00:00:00 +0000

For AAIR the planning question is how much of the risk method you already own, because 79% of the exam is risk discipline applied to AI. ISACA’s three domains are AI Risk Program Management (42%), AI Risk Governance and Framework Integration (37%) and AI Life Cycle Risk Management (21%), across 90 questions. ISACA’s AAIR pages do not publish the duration or passing score, so any figure quoted elsewhere should be checked. The eligibility list is the broadest of ISACA’s AI credentials, which means the candidates who sit it come from very different backgrounds.

AAISM study scenarios: three manager profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AAISM depends on whether your security-management experience comes with technical depth, because the largest domain is the technical one. The exam has 90 questions in three domains: AI Technologies and Controls (38%), AI Governance and Program Management (31%) and AI Risk Management (31%). ISACA’s candidate guide gives 150 minutes and a pass mark of 450 on a scaled 200 to 800 range. Every candidate must hold an active CISM or CISSP, so the profiles differ by which of those two they carry and what the day job looks like.

AB-100 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you study for AB-100 depends mostly on which associate-level certification you already hold, because the exam is the same for everyone: 700 out of 1000 to pass, with Deploy worth 40 to 45 percent of the marks and Plan and Design worth 25 to 30 percent each. The three profiles below show where the hours go for an associate holder on Dynamics 365 or Power Platform, for an AB-620 holder, and for a consultant with no associate certification yet.

AB-620 study paths for three backgrounds

Fri, 09 Oct 2026 00:00:00 +0000

For AB-620 the question is not how much you know about AI but how much of Copilot Studio you have already wired up, because the largest domain, Integrate and extend agents (40 to 45 percent), is all mechanisms you must have configured at least once. The exam lasts 120 minutes and needs 700 out of 1000. The three profiles below cover a Power Platform maker, a professional developer and a builder who already runs agents.

AB-730: three example study profiles

Fri, 09 Oct 2026 00:00:00 +0000

For AB-730 the main question is whether your preparation matches the outline in force from October 20, 2026, which has four skill areas and a new agents and Copilot Cowork domain. The exam is 45 minutes with a passing score of 700, and the plan on this site assumes three to five hours a week. The three profiles below start from three common situations and show where the hours go.

AB-731 study scenarios for three leaders

Fri, 09 Oct 2026 00:00:00 +0000

AB-731 rewards leaders who can choose and justify, and its weights are lopsided: business value of generative AI and Microsoft’s AI apps and services are each 35 to 40 percent, while implementation and adoption is 20 to 25 percent. The exam lasts 45 minutes with a passing score of 700. The three profiles below start from a Copilot rollout lead, a technical professional and a manager new to AI.

AB-900 study scenarios by background

Fri, 09 Oct 2026 00:00:00 +0000

The name suggests a Copilot exam, but the largest AB-900 domain is data protection and governance at 35 to 40 percent, ahead of core Microsoft 365 features (30 to 35 percent) and Copilot and agent administration (25 to 30 percent). The exam lasts 45 minutes with a passing score of 700. The three profiles below show how that weighting shifts study time for an experienced admin, a former MS-900 candidate and a newcomer.

Agentforce study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for the Agentforce Specialist exam depends on which of its six sections you have already touched in a real org. The paper is 60 scored questions in 105 minutes with a flat 72% pass mark (44 correct), and AI Agents alone is 35%, so the biggest question for any background is how fast you can get to a working agent, not how much theory you can read.

AI-103 study scenarios: three example profiles

Fri, 09 Oct 2026 00:00:00 +0000

How to study for AI-103 depends on how much of the older Azure AI services material you already carry. The exam is 120 minutes with a scaled pass mark of 700 out of 1000, and its two largest skill areas (generative AI and agents at 30 to 35%, plan and manage at 25 to 30%) together exceed half the score, so every background should front-load those two and treat the rest as consolidation.

AI-200 study scenarios: three example profiles

Fri, 09 Oct 2026 00:00:00 +0000

The right way to study for AI-200 depends on which half of its title you already know: the cloud developer half or the AI half. The exam is 120 minutes, scaled to 700 out of 1000, with four domains that are close in size, and the largest one, AI solutions on Azure data management services (25 to 30%), is the one with the least overlap with the exam it replaced, AZ-204.

AI-300 study scenarios: example profiles

Fri, 09 Oct 2026 00:00:00 +0000

How to study for AI-300 depends on which of its two halves you already work in. The exam is 120 minutes, English only, with a scaled pass mark of 700 out of 1000, and it splits into Azure Machine Learning (the first two domains, 40 to 50%) and generative AI operations in Microsoft Foundry (the last three, 40 to 55%). Nobody passes on one half alone, so each profile below is really a question of which half is the gap.

AI-901 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How to study for AI-901 depends mostly on whether you can already read Python. It is a fundamentals exam in name, with a scaled 700-out-of-1000 pass mark and only two skill areas, but 55 to 60% of the score sits in implementing AI solutions with Microsoft Foundry and Microsoft’s audience profile assumes Python syntax and familiarity with REST APIs, SDKs and CLIs.

AIF-C01 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AIF-C01 depends mostly on which of two blocks you already know: generative AI and foundation models, which together are 52% of the scored content (24% plus 28%), or the AWS service vocabulary that the whole exam leans on. The exam is compensatory, so you pass on the total rather than per domain, and AWS lists coding models, feature engineering and hyperparameter tuning as out of scope.

AZ-104 study scenarios: three starting points

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AZ-104 depends on whether your gap is Azure vocabulary, administration habits or both. The exam covers five skill areas (identity and governance 20 to 25%, storage 15 to 20%, compute 20 to 25%, networking 15 to 20%, monitoring 10 to 15%), runs 100 minutes, and expects you to read PowerShell, CLI, ARM and Bicep snippets, not just choose a service.

AZ-305 study scenarios: three design profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AZ-305 depends on which kind of decision you are used to making. The exam asks which service meets a requirement, not how to configure it, across four domains: infrastructure solutions 30 to 35%, identity, governance and monitoring 25 to 30%, data storage 20 to 25% and business continuity 15 to 20%. Microsoft does not publish the duration or question count on the exam page, so confirm both when booking.

AZ-400 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AZ-400 depends on one number and one requirement: build and release pipelines are 50 to 55% of the exam, and Microsoft’s audience profile expects experience with both GitHub and Azure DevOps. A candidate who is strong on one product and thin on pipelines, or strong on pipelines in one product only, fails for different reasons, so the plan differs.

AZ-802 study scenarios: three example paths

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AZ-802 depends on which half of “hybrid” you already know. The exam spans seven domains: Active Directory Domain Services 20 to 25%, storage and file services 15 to 20%, monitoring and troubleshooting 15 to 20%, and four domains at 10 to 15% each. It runs 120 minutes, is offered in English only, and has no official practice assessment yet.

AZ-900 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for AZ-900 depends on which of its three skill areas already feels familiar. The exam is 45 minutes, needs 700 out of 1000, and spreads its weight as 25-30% cloud concepts, 35-40% Azure architecture and services, and 30-35% management and governance. The same four-week plan fits most people, but the hours inside it move depending on where you start.

CY0-001 scenarios: three study profiles

Fri, 09 Oct 2026 00:00:00 +0000

The way to study for CY0-001 is to find which of its four domains your current job already covers and spend the saved time on the one that is new. The exam is 60 minutes, up to 60 questions, and a score of 600 on a 100-900 scale. Securing AI systems carries 40%, AI-assisted security 24%, AI governance, risk and compliance 19%, and basic AI concepts 17%. CompTIA recommends 3-4 years in IT with about two years of hands-on security, so every profile below starts from real security work.

Data Engineer Associate: study scenarios

Fri, 09 Oct 2026 00:00:00 +0000

Studying for the Databricks Data Engineer Associate comes down to finding which of its seven sections your background does not touch. The exam has 45 scored multiple-choice questions in 90 minutes, no section above 22%, and an official focus on introductory data engineering tasks. Because the weights are flat, a strong area earns little spare credit, and a blind spot is expensive.

Data Engineer Professional study scenarios

Fri, 09 Oct 2026 00:00:00 +0000

The right way to study for the Databricks Data Engineer Professional depends on how much of its nine sections you already operate. The exam has 60 scored multiple-choice questions in 120 minutes, and the outline changed on 9 October 2026: developing code is 23%, cost and performance 15%, and the remaining seven sections run from 5% to 12%. Because the paper has no dominant area, the main planning question is which small sections you would otherwise give nothing.

Databricks Data Analyst study scenarios

Fri, 09 Oct 2026 00:00:00 +0000

How to study for the Databricks Data Analyst Associate depends on whether your gap is the platform, the dashboard tooling or SQL itself. The exam is 45 scored multiple-choice questions in 90 minutes, English only, with nine sections where the largest is executing queries at 20%. Databricks recommends six or more months of hands-on analysis, and the sections that most often surprise experienced analysts are Genie spaces (12%) and the platform and governance sections.

DP-600 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

To study for DP-600 well, match your background to the outline’s three domains: Prepare data is 45 to 50%, while Maintain a data analytics solution and Implement and manage semantic models are 25 to 30% each. The exam runs 100 minutes with a passing score of 700 on a 1 to 1000 scale, and the English outline changes on October 19, 2026.

DP-700 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for DP-700 depends on which of the exam’s three equal parts you already own. Each skill area carries 30 to 35% of a 100-minute exam with a 700 out of 1000 passing score, so nobody can lean on one strength. The practical question is which third of the outline is thinnest in your own work.

DP-750 example study profiles for Databricks

Fri, 09 Oct 2026 00:00:00 +0000

Studying for DP-750 well means spending hours where the weights are, then correcting for what your job never touches. Two domains carry 30 to 35% each (preparing data, and deploying and maintaining pipelines) and two carry 15 to 20% each (environment setup, and Unity Catalog security and governance). The exam lasts 120 minutes and needs 700 out of 1000.

DP-800 study profiles: three SQL backgrounds

Fri, 09 Oct 2026 00:00:00 +0000

For DP-800, the right study route depends on whether your daily work is T-SQL, application code or database operations, because the exam blends all three. Two domains are weighted at 35 to 40% (design and develop, and secure, optimize and deploy) and the AI domain at 25 to 30%. The exam runs 120 minutes with a 700 out of 1000 passing score.

GenAI Engineer study scenarios: 3 profiles

Fri, 09 Oct 2026 00:00:00 +0000

The way to study for the Databricks Generative AI Engineer Associate exam depends on which of its six sections you already work in, because the paper is 45 scored multiple choice questions in 90 minutes and Application Development alone carries 30% of them. Databricks publishes no passing score, so the useful question is where your own weak sections sit relative to the weights.

ML Associate study scenarios, three profiles

Fri, 09 Oct 2026 00:00:00 +0000

Study for the Databricks Machine Learning Associate exam by checking your background against the 38% platform section first: it is the largest of four sections on a 48-question, 90-minute paper, and it tests MLflow, AutoML and Unity Catalog rather than modelling theory. A candidate strong in modelling and weak on the platform loses the most marks.

ML Professional: study scenarios by profile

Fri, 09 Oct 2026 00:00:00 +0000

For the Databricks Machine Learning Professional exam, study for the 44/44/12 split rather than for “production” in general: Model Development and ML Ops are 44% each, Model Deployment is 12%, and the 59 scored questions run for 120 minutes in English only. Databricks recommends 1+ years of hands-on experience, so the real question is which half of the paper your experience actually covers.

MLA-C02 study scenarios and example profiles

Fri, 09 Oct 2026 00:00:00 +0000

How to study for MLA-C02 depends on which half of “ML engineer” you already do. The exam is flat across four domains: Data Preparation for ML and AI at 28%, and model development, deployment and orchestration, and operating, monitoring and securing at 24% each. The passing score is 720 on a 100 to 1,000 scale, scoring is compensatory, and AWS expects about a year of SageMaker AI and Bedrock experience.

N10-009 study paths for three backgrounds

Fri, 09 Oct 2026 00:00:00 +0000

For Network+ N10-009, the right preparation depends less on how much you know than on whether you can subnet quickly and reason through a fault. The exam gives at most 90 questions in 90 minutes, passes at 720 on a scale of 100 to 900, and weights troubleshooting highest at 24%, ahead of networking concepts at 23%.

PL-300 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

How you should study for PL-300 depends on which of its four domains your daily work already covers. Three of them (Prepare the data, Model the data, Visualize and analyze the data) are each worth 25–30%, and Manage and secure Power BI is worth 15–20%, so a background that is strong in one domain still leaves roughly two thirds of the exam to learn or verify. The 100-minute exam needs 700 out of 1000.

SC-200 scenarios: which study route fits

Fri, 09 Oct 2026 00:00:00 +0000

To study for SC-200 well, work out whether your gaps are in KQL, in Microsoft Sentinel and Defender configuration, or in security operations itself. The exam has three domains: Manage a security operations environment (40–45%), Respond to security incidents (35–40%) and Perform threat hunting (20–25%), and it runs 100 minutes with a passing score of 700 out of 1000. KQL appears in all three.

SC-300 study scenarios: choose your route

Fri, 09 Oct 2026 00:00:00 +0000

For SC-300, the right study route depends on which of the exam’s four domains you already operate: users and hybrid identity, authentication and Conditional Access, workload identities, or governance. The weights are close, Authentication and access management is largest at 25–30%, and the other three sit at 20–25% each. The exam is 100 minutes and needs 700 out of 1000, and the English outline changes on October 28, 2026.

SC-401 study scenarios: three planning profiles

Fri, 09 Oct 2026 00:00:00 +0000

How to study for SC-401 depends on how much Microsoft Purview you have actually configured, because the exam is three nearly equal domains (Implement information protection, Implement data loss prevention and retention, Manage risks, alerts, and activities, each 30–35%) and rewards people who have clicked through the settings. It lasts 100 minutes and needs 700 out of 1000.

SC-500 study scenarios: where your hours go

Fri, 09 Oct 2026 00:00:00 +0000

For SC-500, plan for coverage rather than depth, because its four domains are nearly equal: Manage identity, access, and governance (20–25%), Secure storage, databases, and networking (25–30%), Secure compute (20–25%) and Manage and monitor security posture (20–25%). The exam lasts 120 minutes and needs 700 out of 1000, and no domain is small enough to skip.

SC-900 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

SC-900 is a 45-minute, describe-level exam with a passing score of 700, and its biggest area is Microsoft security solutions at 35 to 40 percent, ahead of Entra at 25 to 30 percent. Your background decides which of the four skill areas you can skim and which one needs most of the hours. The three profiles below show how that plays out for an IT practitioner, a business-side reader and a complete newcomer.

SEE Part 1 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

Study time for SEE Part 1 should follow the 85 scored questions, not the 100 on the paper: Income and Assets and Deductions and Credits carry 17 each, Taxation 15, Preliminary Work 14, and Advising and Specialized Returns 11 each. The study plan assumes eight weeks at 8 to 10 hours a week for someone who knows basic tax, and your background moves that figure in either direction. The three profiles below show how.

SEE Part 2 study scenarios: example profiles

Fri, 09 Oct 2026 00:00:00 +0000

SEE Part 2 puts 37 of its 85 scored questions in Business Tax Preparation, 30 in Business Entities and Considerations and 18 in Specialized Returns and Taxpayers, so how you study depends on which of those you already use at work. The study plan assumes six weeks at 8 to 10 hours for someone who prepares returns, and eight weeks if business returns are new. The three profiles below show where that budget moves.

SEE Part 3 study scenarios: profiles

Fri, 09 Oct 2026 00:00:00 +0000

SEE Part 3 is flatter than the other two parts, with 26, 25, 20 and 14 scored questions across Practices and Procedures, Representation, Specific Areas of Representation and the Filing Process, so no domain can be skipped. The study plan assumes six weeks at 8 to 10 hours for someone who has not represented clients, and four weeks for someone who has. The three profiles below show how that range is used.

Spark Developer study scenarios: 3 profiles

Fri, 09 Oct 2026 00:00:00 +0000

The way to study for the Databricks Associate Developer for Apache Spark exam depends on whether you can already read PySpark and reason about shuffles, because the paper is 45 Python-only questions in 90 minutes and half of it is the DataFrame API (30%) plus Spark SQL (20%). Architecture adds another 20%, which most associate exams do not demand.

SY0-801 study scenarios: three profiles

Fri, 09 Oct 2026 00:00:00 +0000

SY0-801 is a maximum of 90 questions in 90 minutes with a passing score of 750 on a 100 to 900 scale, and Security Operations at 27 percent and Threats, Vulnerabilities, and Attacks at 24 percent are about half of it. CompTIA lists the V8 launch as expected on or around November 17, 2026, so the practical question is how much of your time goes where while you wait. The three profiles below cover a candidate with SY0-701 material, a help desk or network technician and a student.

AAA authentication, authorization, accounting

Fri, 09 Oct 2026 00:00:00 +0000

AAA is three separate questions asked about every access attempt. Authentication: who are you? Authorization: what may you do? Accounting: what did you do? The decision rule for exams: authentication always comes first and proves identity, authorization is decided after it and limits actions, and accounting records both afterwards. If a question describes proving a claim, it is authentication; if it describes permissions, it is authorization; if it describes logs, usage records or an audit trail, it is accounting.

Access reviews and entitlement management

Fri, 09 Oct 2026 00:00:00 +0000

Entitlement management and access reviews are two halves of Microsoft Entra ID Governance. Entitlement management controls how access is requested, approved, granted and expired, using access packages. Access reviews ask people periodically whether existing access is still needed and can remove it. The decision rule: if the scenario is about getting access in the right way (request, approval, time limit, partner users), the answer is entitlement management; if it is about checking access that already exists (stale members, leaving guests, too many admins), the answer is an access review. Many good answers use both: an access package whose policy includes a recurring review.

ACL practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free ACL practice questions with explained answers on wildcard masks, rule order, the implicit deny, placement and verification. An access control list (ACL) is an ordered list of permit and deny statements that a Cisco router applies to packets entering or leaving an interface. ACLs are tested on the CCNA 200-301 and appear conceptually on CompTIA Network+ N10-009. The questions assume IPv4 ACLs on Cisco IOS.

How to use these questions

Attempt each question before opening the answer. For every list, read top to bottom and stop at the first match, and remember that an implicit deny closes every ACL. Note the constraint words: “inbound”, “closest to”, “only” and “originated by the router” decide several answers. The rules behind them are in ACLs explained.

ACLs explained: wildcard masks and order

Fri, 09 Oct 2026 00:00:00 +0000

Standard or extended: the choice that starts every ACL question

An access control list (ACL) is an ordered list of permit and deny statements that a router applies to packets on an interface, in one direction. A standard ACL matches only the source address; an extended ACL matches source, destination, protocol and port numbers. The decision rule: if the requirement names only who the traffic comes from, use standard; the moment it mentions a destination, a protocol or a port, you need extended. Whichever you use, the router checks entries top to bottom, the first match decides, and an invisible deny any waits at the end.

Administrative distance and the routing table

Fri, 09 Oct 2026 00:00:00 +0000

The lookup order in one paragraph

A router chooses a route in two separate steps. First, for each destination prefix it keeps only the best source: if several routing sources offer the same prefix and mask, the one with the lowest administrative distance (AD) is installed in the routing table, and the metric only breaks ties inside one protocol. Second, when a packet arrives, the router forwards it using the longest prefix match among the installed routes. The decision rule for exam questions: find the most specific prefix that contains the destination address, and only if two sources offer that exact prefix compare AD, then metric. AD never overrides a longer prefix.

AI agents explained: tools, memory, guardrails

Fri, 09 Oct 2026 00:00:00 +0000

An AI agent is a language model that is given a goal, instructions and a set of tools, and decides for itself which step to take next, looping until the task is done. A plain chatbot call answers once; an agent can look something up, call an API, check the result and try again. The decision rule for exams: when the requirement is to know something, use retrieval; when it is to do something, use a tool the agent calls; and when an action cannot be undone, put a human approval step or a deterministic rule in front of it.

AI governance: NIST AI RMF, ISO 42001, EU AI Act

Fri, 09 Oct 2026 00:00:00 +0000

The three names that dominate AI governance questions are different kinds of thing. The NIST AI Risk Management Framework is a voluntary framework that tells an organisation how to think about AI risk. ISO/IEC 42001 is a management system standard that an auditor can certify. The EU AI Act is binding law that sorts AI systems into risk tiers and attaches legal duties to each tier. The decision rule: if a question asks for a way to structure risk work, think NIST; for something to be audited and certified, think ISO 42001; for something that is mandatory and depends on how risky the use case is, think the EU AI Act.

Ansible vs Terraform: which tool for what

Fri, 09 Oct 2026 00:00:00 +0000

The one-line rule

Terraform creates and changes infrastructure (networks, virtual machines, load balancers, DNS records) from a declared desired state, and it remembers what it built in a state file. Ansible configures what already exists (packages, files, users, services, device settings) by running tasks from a playbook on each managed host, without installing an agent. The decision rule for an exam: if the scenario is about making resources appear, pick Terraform; if it is about changing the settings on machines or devices that are already there, pick Ansible. Both can do a little of the other’s job, which is exactly why exam questions are built on the difference in their design, not on what is theoretically possible.

App Service plans: tiers, scaling and sharing

Fri, 09 Oct 2026 00:00:00 +0000

An Azure App Service plan is the set of compute resources (virtual machines) that one or more App Service apps run on. The plan, not the app, decides the region, operating system, VM size, instance count and pricing tier, and it is the thing you pay for. The decision rule for exams: features such as autoscale and deployment slots are unlocked by the tier of the plan, capacity is changed by scaling the plan, and every app in a plan shares the same VMs and scales together; to give an app its own capacity, move it to its own plan.

ARM templates vs Bicep vs Terraform

Fri, 09 Oct 2026 00:00:00 +0000

ARM templates, Bicep and Terraform are all declarative infrastructure as code: you describe the end state and the tool works out the steps. ARM templates are the JSON format Azure Resource Manager reads, Bicep is a shorter language that compiles to that same JSON, and Terraform is a separate multi-cloud tool with its own state file. The decision rule: if the scenario is Azure only and says nothing about other clouds or Terraform, the answer is Bicep; if it mentions several clouds, an existing Terraform codebase or a provider for a non-Azure service, the answer is Terraform; ARM JSON is the answer only when the question is about the template format itself.

Auto Loader: incremental file ingestion

Fri, 09 Oct 2026 00:00:00 +0000

Auto Loader is the Databricks feature for incrementally ingesting files that land in cloud storage: it remembers which files it has already processed and loads only the new ones. The decision rule for exams: if new files keep arriving in a storage location and each must be loaded exactly once, with the schema allowed to change, use Auto Loader (the cloudFiles source). If the load is a simple scheduled SQL statement over a folder, COPY INTO also works; if it is a database or SaaS source rather than files, the answer is a Lakeflow Connect connector, not Auto Loader.

Availability sets vs availability zones

Fri, 09 Oct 2026 00:00:00 +0000

An availability set spreads virtual machines across fault domains and update domains inside one datacenter, while availability zones spread them across physically separate datacenters in one region. The decision rule is to read the failure the scenario names: a rack, power or maintenance problem is covered by an availability set, the loss of a whole datacenter needs zones, and the loss of a whole region needs a second region, which neither option provides.

AWS AI practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free AWS AI services practice questions with explained answers, written as short business scenarios. AWS AI services are pre-trained, API-based services for images, documents, text, speech, conversation and recommendations that need no model training. They are tested in the AWS Certified AI Practitioner (AIF-C01) and the AWS Certified Machine Learning Engineer Associate (MLA-C02) exams. All questions are original and written against the official outlines.

How to use these questions

Attempt each question before opening the answer. Name the input data type and the wanted output first, then match the service. Note the constraint word in the stem (“least development effort”, “no ML expertise”, “live”, “already recorded”), because it usually separates the right answer from a plausible one. The AWS AI services topic guide has the input-to-output cheat sheet and the list of services closed to new customers. Difficulty rises across the three sections: the first five questions match one input to one output, the middle five separate services that look alike, and the last five add constraints such as custom categories, availability and fewest services.

AWS AI services: match the service to the task

Fri, 09 Oct 2026 00:00:00 +0000

AWS AI services are pre-trained, API-based services for common problems (images, documents, text, speech, conversation, recommendations) that need no model training. The decision rule: name the input and the output, and pick the service whose job that is; build a custom model in SageMaker AI only when no managed service covers the task or the exam scenario demands a custom algorithm. Most exam questions on this topic are one-line matching problems, so the skill is knowing the input and output of each service cold.

AWS IAM basics: roles, policies, evaluation

Fri, 09 Oct 2026 00:00:00 +0000

AWS Identity and Access Management (IAM) decides who can do what to which AWS resource, using JSON policies that are evaluated on every request. The decision rule for exams: give workloads an IAM role that supplies temporary credentials, grant only the actions and resources the task needs, and remember that an explicit deny always wins. If an answer option stores an access key in code or in a container, it is almost never the right one.

Azure Backup vs Site Recovery explained

Fri, 09 Oct 2026 00:00:00 +0000

Azure Backup keeps point-in-time copies of data so you can restore something that was lost or damaged, while Azure Site Recovery continuously replicates running machines to another location so you can fail over when a site is down. The decision rule is that “restore data from before the problem” means Backup, and “keep the workload running elsewhere” means Site Recovery. Large designs use both, because replication copies corruption and deletion just as faithfully as good data.

Azure Firewall vs NSG vs WAF: which to pick

Fri, 09 Oct 2026 00:00:00 +0000

A network security group (NSG) filters packets by address, port and protocol on a subnet or network interface. Azure Firewall is a central, stateful firewall service that adds FQDN filtering, threat intelligence and (in Premium) intrusion detection. A Web Application Firewall (WAF) inspects HTTP and HTTPS requests for attacks such as SQL injection and cross-site scripting. The decision rule: name what has to be inspected. Ports and IP ranges means NSG; outbound domain names or one central control point means Azure Firewall; web attacks means WAF. They do not replace each other, and most real designs use all three.

Azure Functions triggers and bindings

Fri, 09 Oct 2026 00:00:00 +0000

In Azure Functions, a trigger is what starts a function and a binding is a declarative connection to another service, either as input (data passed in) or output (data written out). The decision rule for exams: every function has exactly one trigger and any number of optional input and output bindings; pick the trigger by what should cause the code to run, and use an output binding when the only job is to write the result somewhere.

Azure governance design for AZ-305

Fri, 09 Oct 2026 00:00:00 +0000

Azure governance design means choosing where controls are applied so that every current and future subscription inherits them. The rule for AZ-305 answers: put the control at the highest scope where it is true for everything beneath it, and use management groups, not repeated per-subscription assignments, to get there. Policy says what is allowed, RBAC says who may act, and the hierarchy decides how far each one reaches.

The scopes and what each one is for

Azure resources sit in four levels of scope, and every management group tree starts at a single root management group:

Azure Key Vault: access, network and recovery

Fri, 09 Oct 2026 00:00:00 +0000

Azure Key Vault is a managed store for secrets (passwords, connection strings), keys (for encryption and signing) and certificates, with Microsoft Entra authentication, auditing and recovery built in. The decision rule for exams: an app should read a secret at run time using a managed identity that holds a data-plane role such as Key Vault Secrets User, never a stored credential; control-plane roles manage the vault but do not read its contents; and soft delete plus purge protection, not backups, are what make a deleted object recoverable.

Azure load balancing options compared

Fri, 09 Oct 2026 00:00:00 +0000

Azure has four load balancing services that exam questions mix freely: Load Balancer (layer 4, regional), Application Gateway (layer 7, regional), Front Door (layer 7, global) and Traffic Manager (DNS-based, global). The decision rule is two questions: is the traffic HTTP(S) that needs URL, host or WAF decisions, and does it need to span regions? Non-HTTP and regional gives Load Balancer, HTTP and regional gives Application Gateway, HTTP and global gives Front Door, and non-HTTP or DNS-only and global gives Traffic Manager.

Azure Monitor and Log Analytics explained

Fri, 09 Oct 2026 00:00:00 +0000

Azure Monitor is the platform that collects, stores and acts on telemetry from Azure resources. It keeps two kinds of data in two different stores: metrics (numbers over time, collected automatically) and logs (records you query with KQL, stored in a Log Analytics workspace). The decision rule for most exam questions: if the data is a number that needs little processing, use metrics; if you need to search, join, retain or correlate records, send logs to a workspace with a diagnostic setting or a data collection rule.

Azure Policy vs RBAC vs resource locks

Fri, 09 Oct 2026 00:00:00 +0000

Azure RBAC decides who may perform an action, Azure Policy decides what resource configurations are allowed regardless of who deploys them, and a resource lock stops deletion or modification for everyone until the lock is removed. The decision rule is to ask which noun the requirement is about: a person or group points to RBAC, a property of the resource such as region, SKU or tag points to Policy, and “must not be deleted” points to a lock.

Azure RBAC practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Azure RBAC practice questions with explained answers, written as short scenarios about roles, scopes and the split between Azure roles and Microsoft Entra roles. Azure role-based access control decides who can do what with Azure resources, while Entra roles govern the directory. The topic is tested in AZ-104, AZ-900 and SC-300, among others.

How to use these questions

Try each question before opening the explanation and commit to one letter. Underline the constraint word in the stem, such as “least privilege”, “without granting access” or “only for this resource group”, because it usually separates the right role from a near miss. Read why each wrong option fails, since the same distractors return on real exams. The guide Azure RBAC vs Microsoft Entra roles holds the theory behind these fifteen.

Azure RBAC vs Microsoft Entra roles

Fri, 09 Oct 2026 00:00:00 +0000

Azure RBAC roles control what you can do with Azure resources (virtual machines, storage, networks). Microsoft Entra roles control what you can do in the directory (users, groups, domains, licences, Conditional Access). The two systems are separate: by default, a Global Administrator has no access to Azure resources, and an Owner of a subscription cannot reset a user’s password. The decision rule for exams: ask what the object being managed is. A resource in a subscription means Azure RBAC; an identity or tenant setting means an Entra role.

Azure Resource Hierarchy Explained

Fri, 09 Oct 2026 00:00:00 +0000

The Azure resource hierarchy has four levels of scope: management groups, subscriptions, resource groups and resources. Settings applied at a higher level, such as role assignments and policy, flow down to everything below it, and they never flow up or sideways. The decision rule is to assign at the highest level that matches the requirement, and no higher: one rule for many subscriptions belongs on a management group, one rule for one application belongs on its resource group.

Azure Storage redundancy practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Azure Storage practice questions with explained answers, written as short scenarios about choosing and operating the redundancy option of a storage account. Redundancy decides how many copies of the data exist and where they sit, from LRS in one datacenter to RA-GZRS across zones and regions. The topic is tested in AZ-900, AZ-104 and AZ-305.

How to use these questions

Answer each question before opening the explanation, and commit to one letter. Underline the constraint word in the stem, such as “datacenter”, “region”, “read during the outage” or “lowest cost”, because it separates neighbouring options. Read why each wrong option fails, since the same distractors return on real exams. The guide Azure Storage redundancy: LRS to RA-GZRS holds the decision table behind the answers.

Azure Storage redundancy: LRS to RA-GZRS

Fri, 09 Oct 2026 00:00:00 +0000

Azure Storage redundancy is the setting that decides how many copies of your data exist and where they sit. The decision rule is to name the failure you must survive, then take the cheapest option that survives it: a datacenter outage needs ZRS, a regional outage needs GRS or GZRS, and reading data during a regional outage without waiting for a failover needs the RA- version of either.

The decision table

Every option keeps three copies in the primary region. The geo options add a second set in the paired region, which is chosen by Azure and cannot be changed.

Azure Storage Types: Blob, Files, Disks, Queue

Fri, 09 Oct 2026 00:00:00 +0000

Azure storage questions ask you to match the shape of the data and the way it is accessed to a service: objects and unstructured data go to Blob Storage, shared file shares to Azure Files, messages to Queue Storage, key-attribute records to Table Storage, and a virtual machine’s own volumes to managed disks. The decision rule is to name the access method first (HTTP object, SMB or NFS share, message, block volume) and only then pick the account type and access tier.

Azure VMs, Availability Sets and Scale Sets

Fri, 09 Oct 2026 00:00:00 +0000

A single Azure virtual machine gives you one instance with full OS control, an availability set spreads a handful of hand-built VMs across fault domains, and a Virtual Machine Scale Set (VMSS) manages a group of VMs that can grow and shrink automatically. The decision rule is to ask whether the number of instances must change on its own: if yes, use a scale set; if the VMs are fixed, different from each other and long-lived, use individual VMs placed across availability zones or, where zones do not exist, in an availability set.

Azure VNets, subnets and peering explained

Fri, 09 Oct 2026 00:00:00 +0000

An Azure virtual network (VNet) is a private address space in one region, a subnet is a slice of that space where resources get their IPs, and peering links two VNets so they talk over Microsoft’s backbone. Two rules decide most exam questions: Azure keeps five addresses in every subnet, and peering is not transitive (A to B and B to C does not give A to C).

Layer 1: the VNet is an address plan

You choose one or more address spaces in CIDR notation, typically from the private ranges, such as 10.20.0.0/16. Nothing outside is reachable yet: a VNet is a boundary, and resources inside it can talk to each other by default. A VNet also lives in one region and one subscription, which is why connecting several of them is a design question and not a setting.

Backup types, RTO and RPO explained

Fri, 09 Oct 2026 00:00:00 +0000

A backup type decides how much data is copied each night and how many pieces you need on restore day. RPO (recovery point objective) is how much data the business can afford to lose, measured as time. RTO (recovery time objective) is how long the service can stay down. The exam rule: RPO sets how often you must back up or replicate, RTO sets how fast your restore method and recovery site must be, and the cheapest design that meets both numbers is the answer.

Batch vs streaming: how to choose

Fri, 09 Oct 2026 00:00:00 +0000

Batch processing handles a bounded set of data in one run, then stops; streaming processes events continuously as they arrive, on data that has no end. The decision rule for exams: choose by the latency the business asks for, not by what sounds modern. If a result is needed once an hour or once a day, batch is simpler and cheaper; if the stem says “as events arrive”, “within seconds” or “near real time”, use a streaming engine; if it says “incrementally, but only run when scheduled”, the middle option (a streaming engine run in batch-like mode) is often the answer.

Bedrock vs SageMaker AI: which one to pick

Fri, 09 Oct 2026 00:00:00 +0000

Amazon Bedrock is a fully managed service that gives you API access to foundation models from several providers; you build the application and AWS runs the models. Amazon SageMaker AI is the platform for building, training and hosting machine learning models yourself, including open foundation models. The decision rule: if calling an existing foundation model through an API is enough, choose Bedrock; if you must train a model, run classical ML, or control the hosting infrastructure, choose SageMaker AI. The two are not rivals. A model trained in SageMaker AI can be imported into Bedrock, and Bedrock models can be evaluated and orchestrated from SageMaker pipelines.

BGP basics: what to know for Network+

Fri, 09 Oct 2026 00:00:00 +0000

BGP in one paragraph

BGP (Border Gateway Protocol, version 4, defined in RFC 4271) is the exterior routing protocol that connects autonomous systems: the separately administered networks of ISPs, large companies and cloud providers that together make up the internet. Where OSPF and EIGRP find the best path inside one organisation by metric, BGP exchanges reachability between organisations and chooses routes by policy and by the list of autonomous systems a route has crossed. The exam rule: inside one network, use an interior protocol; between two autonomous systems, such as a company and its ISP, the answer is BGP. Network+ N10-009 names BGP but only at recognition depth; the CCNA objectives do not list it.

Bicep, ARM and Terraform practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free infrastructure as code practice questions with explained answers, written as short Azure scenarios about ARM templates, Bicep and Terraform. Infrastructure as code means describing the desired end state of resources in files that a tool deploys repeatably. The topic is tested in AZ-104, AZ-400 and AZ-900, with a smaller share in other Microsoft exams.

How to use these questions

Try each question before opening the explanation and commit to one letter. Underline the constraint word in the stem, such as “no state file”, “preview”, “delete resources removed from the file” or “several clouds”, because that word usually separates Bicep, ARM JSON and Terraform. Read why each wrong option fails, since the same distractors return on real exams. The guide ARM templates vs Bicep vs Terraform holds the theory behind these fifteen.

Business entity types compared for US tax

Fri, 09 Oct 2026 00:00:00 +0000

For federal income tax, a business is taxed in one of two ways: it is either a pass-through, where the income lands on the owners’ own returns, or a separate taxpayer, as a C corporation. Sole proprietorships, partnerships and S corporations are pass-throughs; a C corporation pays its own tax. An LLC has no tax identity of its own, because it takes the treatment of whichever type it defaults to or elects. The decision rule: identify the legal form, apply the default classification, then check whether an election (Form 8832 or Form 2553) changed it. This guide describes the rules as published by the IRS for 2025 returns; the sources are linked where each rule appears.

Cabling and Connectors: Cat6, Fiber, SFP

Fri, 09 Oct 2026 00:00:00 +0000

The one-line rule for choosing a cable

Pick the cable from the distance and speed the scenario gives you: twisted pair copper (UTP) for runs up to 100 m, multimode fiber for building and campus runs, single-mode fiber for long distances, and direct attach copper (DAC) for a few metres between devices in the same or neighbouring racks. Then pick the connector that matches the port: RJ45 for twisted pair, LC or SC for fiber, F-type or BNC for coaxial. Most exam questions are solved by those two steps.

Capital gains and losses: netting and rates

Fri, 09 Oct 2026 00:00:00 +0000

A capital gain or loss is the difference between what a taxpayer receives for a capital asset and the asset’s adjusted basis. The decision rule for exam questions: sort every sale into short-term (held one year or less) or long-term (held more than one year), net within each bucket, net the two results against each other, and only then apply the rate or the loss limit. This guide uses tax year 2025 and the IRS sources Publication 550 (2025) and the 2025 Instructions for Schedule D. It is exam preparation, not tax advice.

Choosing an Azure compute service

Fri, 09 Oct 2026 00:00:00 +0000

Pick the Azure compute service that takes over the most work you do not need to control. Virtual Machines for full OS control or lift and shift, App Service for web apps and APIs, Functions for event-triggered code, Container Apps for containers without Kubernetes, and AKS when you need the Kubernetes API itself. The decision rule is a ladder: start at the managed end and move toward more control only when a requirement forces it, such as a custom OS, a Kubernetes-specific tool, or a legacy installer.

CI/CD and Git practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free CI/CD and Git practice questions with explained answers, written for anyone preparing for a Microsoft DevOps exam. Continuous integration and continuous delivery (CI/CD) automate building, testing and releasing code from a Git repository, and the exam asks which control, command or strategy fits a described situation. The topic is tested in the source control and build and release domains of AZ-400. All questions are original and use Azure DevOps and GitHub terminology.

CI/CD explained: pipeline stages and gates

Fri, 09 Oct 2026 00:00:00 +0000

CI/CD is automation of the path from a code commit to running software. Continuous integration builds and tests every change merged into the shared branch; continuous delivery keeps every build releasable and deploys it through environments with a human deciding when to go to production; continuous deployment removes that human decision. The deciding question on an exam is always the same: is there a manual step before production?

Three terms, one line each

Continuous integration (CI). Microsoft defines it as automatically building and testing code every time a team member commits to version control. A commit to the main branch triggers a build, tests and validation. The point is to merge often, so conflicts and bugs surface early.
Continuous delivery (CD). Microsoft describes it as automating build, test, configuration and deployment from a build to a production environment through a release pipeline. Successive environments support progressively longer-running integration, load and user acceptance tests.
Continuous deployment. The common usage of the term is delivery without the final manual approval: every change that passes all automated checks goes live. Microsoft’s pages use “continuous delivery” for the whole pipeline and allow an optional manual approval before the next ring, so the strict distinction is a convention that exams apply, not a Microsoft definition. Read the stem for whether a person approves production.

The initials CD are therefore ambiguous. When an exam writes “CD”, check whether the stem says approval is required.

Circular 230 basics for the SEE exam

Fri, 09 Oct 2026 00:00:00 +0000

Circular 230 is the Treasury regulation, 31 CFR Part 10, that governs who may practise before the IRS and how they must behave. Its decision rule for exam questions: first ask whether the person is a practitioner or a paid return preparer, then match the situation to one duty section (§10.20 to §10.37) and, if the duty is broken, to a sanction under §10.50. This guide is exam preparation, not advice on a real engagement.

Common attack types: tell them apart on exams

Fri, 09 Oct 2026 00:00:00 +0000

Attack-type questions describe what the attacker did and ask you to name the attack, or name the control that stops it. The decision rule: ignore the story and find the one mechanism, meaning what was abused (a person, a password, an input field, a session, a network position, a resource limit), then match it to the name. Most wrong answers are lookalike attacks that abuse something different.

Clue-to-attack lookup

Read the scenario for the thing that was tricked or exhausted.

Common ports cheat sheet: secure vs insecure

Fri, 09 Oct 2026 00:00:00 +0000

The pairing rule

A port number identifies a service on a host, and the exam question usually gives you a protocol and wants the number, or gives you an insecure protocol and wants the secure replacement. The decision rule: if a protocol sends credentials or data in clear text, there is an encrypted twin, and the twin is the right answer whenever the question mentions encrypted, secure or compliance. All numbers below were checked against the IANA service name and port number registry; protocol behaviour is from the RFCs named in each section.

Conditional Access explained for exams

Fri, 09 Oct 2026 00:00:00 +0000

Conditional Access is the policy engine in Microsoft Entra that decides, after a user has passed the first sign-in step, whether to block access, allow it, or allow it only if extra requirements are met (MFA, a compliant device, a password change). Each policy is an if-then rule. The decision rule for exams: if the scenario says access should depend on a condition (location, device state, risk, which app), the answer is Conditional Access; if it says a role should be held only temporarily, it is Privileged Identity Management instead.

Conditional Access practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Conditional Access practice questions with explained answers, written as short scenarios in the style of Microsoft identity exams. Conditional Access is the Microsoft Entra policy engine that decides, after first-factor sign-in, whether to block access, allow it, or allow it only with extra requirements such as MFA or a compliant device. It is tested in SC-300, SC-900 and AZ-900, among others.

How to use these questions

Answer each question before opening the explanation, and commit to one letter. Note the constraint word in the stem (“only when”, “without affecting”, “block”, “least”), because that word usually decides between two plausible options. Read the explanation of every option, not only the right one, since the wrong options are the traps the exams reuse. The guide Conditional Access explained covers the theory behind all fifteen.

Copilot vs Copilot Studio vs Foundry vs GitHub

Fri, 09 Oct 2026 00:00:00 +0000

Four Microsoft products share the word Copilot or sit next to it, and exams test whether you can tell them apart. Microsoft 365 Copilot is the assistant you use inside Word, Excel, Outlook and Teams. Copilot Studio is the low-code tool for building agents. Microsoft Foundry is the developer platform for building AI apps and agents with code and choosing models. GitHub Copilot is the AI coding assistant for software developers. The decision rule: ask who the user is and whether they are consuming, configuring or coding.

Cryptography basics for security exams

Fri, 09 Oct 2026 00:00:00 +0000

Cryptography on a security exam is a matching game: each tool solves one problem. Symmetric encryption (one shared key) protects the confidentiality of bulk data. Asymmetric encryption (a public and private key pair) solves key exchange and digital signatures. Hashing (one-way, no key) proves integrity. PKI (certificates and certificate authorities) makes people trust that a public key really belongs to who it claims. The decision rule: name the property the scenario needs, such as confidentiality, integrity, authentication or non-repudiation, then pick the tool that provides it.

Cryptography practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free cryptography practice questions with explained answers, written as scenarios. Cryptography protects confidentiality, integrity and authenticity using ciphers, hashes, digital signatures and certificates, and exam questions ask which tool fits which goal. The topic is tested in the CompTIA Security+ V8 exam (SY0-801), objective 1.3 on cryptographic solutions.

How to use these questions

Attempt every question before reading the explanation. Identify the goal first (confidentiality, integrity, authentication or non-repudiation), then pick the primitive that provides it. Wrong options are explained, because the usual trap is a real technique applied to the wrong goal. The cryptography basics guide explains each concept in depth.

Data classification in Microsoft Purview

Fri, 09 Oct 2026 00:00:00 +0000

Data classification in Microsoft Purview is how the platform recognizes which content is sensitive, so that labels, DLP, retention and insider risk policies have something to act on. Learn contrasts three ways of identifying items: manually by users, by automated pattern matching (sensitive information types), and by machine learning (trainable classifiers). The decision rule: if the content has a recognizable pattern, use a sensitive information type; if it is an exact known value, use exact data match; if it is a form that varies, use document fingerprinting; if it is recognizable only by meaning, use a trainable classifier.

Databricks compute types: all-purpose, job, SQL

Fri, 09 Oct 2026 00:00:00 +0000

Databricks offers three families of compute: serverless compute, classic compute (all-purpose and job clusters that you configure), and SQL warehouses. The decision rule for exams: SQL queries and BI go to a SQL warehouse, unattended scheduled work goes to serverless or job compute, and all-purpose compute is for interactive development. If a question puts an all-purpose cluster behind a production job, it is usually the wrong answer.

Cluster or compute?

Older material says “cluster”. Current documentation says “compute” in the UI, while the APIs, the Clusters API and settings such as cluster_id keep the old word. Azure Databricks documentation lists the families as serverless compute, classic compute and SQL warehouses. Within classic compute, the configuration reference still talks about all-purpose and job compute, and job compute can be configured as classic jobs compute. The access modes were renamed too: shared became standard and single user became dedicated.

DAX basics: filter context and CALCULATE

Fri, 09 Oct 2026 00:00:00 +0000

DAX is the formula language of Power BI, Analysis Services and Fabric semantic models, and nearly everything in it depends on context: the set of rows a formula can see at the moment it runs. Row context means “the current row” and exists in calculated columns and inside iterator functions. Filter context means “the rows that survive the filters on this report cell” and exists for every measure. The decision rule: when a question asks you to change which rows a measure sees, the answer is almost always CALCULATE, because it is the only function whose job is to rewrite filter context.

DAX practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free DAX practice questions with explained answers, written as short scenarios on measures, filter context, CALCULATE and iterators. DAX is the formula language of Power BI, Analysis Services and Fabric semantic models, and each question states its tables, relationships and sample values so that the result can be worked out by hand. The topic is tested in PL-300 and DP-600.

How to use these questions

Read the model description first and sketch the relationship direction. Then work out which filters are active in the cell the stem asks about, and commit to one letter before opening the explanation. Underline the constraint word, such as “respond to slicers”, “percent of selected”, “each product” or “contiguous”, because it usually decides between a measure, an iterator and a filter modifier. The guide DAX basics: filter context and CALCULATE explains the mechanism behind the answers.

Defender for Endpoint: plans, ASR, response

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Defender for Endpoint is Microsoft’s endpoint security platform: it prevents, detects, investigates and responds to threats on Windows, macOS, Linux, Android and iOS devices, and feeds its signals into the Microsoft Defender portal and Defender XDR. The decision rule for exams: Plan 1 is prevention plus manual response, Plan 2 adds detection and automated investigation (EDR, hunting, vulnerability management), and every setting that targets “some devices” runs through device groups.

Defender XDR vs Defender for Cloud

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Defender XDR correlates alerts about users, devices, email, identities and SaaS apps into incidents, and Microsoft Defender for Cloud protects and assesses Azure, AWS and Google Cloud resources. The decision rule: if the thing being attacked is a person, a laptop, a mailbox or a SaaS app, the answer is in the Defender XDR family; if it is a subscription, a virtual machine, a storage account, a container cluster or a database in a cloud, the answer is Defender for Cloud. They overlap at the edges, and exams are built on those edges.

Delta Lake and Spark practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Delta Lake and Spark practice questions with explained answers, written for anyone preparing for a Databricks certification. Delta Lake is the open table format that adds a transaction log, ACID transactions and time travel on top of Parquet files in cloud storage, and Spark is the engine that reads and writes it. Both are tested in the Databricks Data Engineer Associate and the Spark Developer Associate exams. All questions are original, use Databricks SQL and PySpark, and assume Unity Catalog where tables are mentioned.

Delta Lake explained: log, ACID, time travel

Fri, 09 Oct 2026 00:00:00 +0000

Delta Lake is an open table format that puts a transaction log on top of ordinary Parquet files in cloud storage, which gives a data lake the behaviour of a database table: atomic writes, consistent reads, schema checks and history. The decision rule for exams: if a scenario needs updates, deletes, merges, safe concurrent writers, or the ability to read an earlier state of the data, the answer is a Delta table, not plain Parquet, CSV or JSON files.

Delta Sharing (now OpenSharing) explained

Fri, 09 Oct 2026 00:00:00 +0000

Delta Sharing is the Databricks open protocol for giving another organization read-only access to live data without copying it. Current Databricks documentation has renamed the feature OpenSharing, so exam material may use either name. The decision rule for exams: a partner who needs ongoing access to your data without an export and without joining your workspace points to Delta Sharing. If the recipient also has a Unity Catalog-enabled Databricks workspace, use Databricks-to-Databricks sharing, and if they are on any other platform, use open sharing.

Depreciation and MACRS: 179, bonus, recovery

Fri, 09 Oct 2026 00:00:00 +0000

Depreciation lets a business deduct the cost of an asset over time instead of all at once. For US tax, three tools apply in a fixed order: the section 179 election, then the special depreciation allowance (bonus), then regular MACRS depreciation on whatever basis is left. The decision rule: pick the asset’s MACRS class first, then check whether 179 or bonus accelerates the deduction, and remember that any depreciation taken reduces basis and can come back as ordinary income on a sale. This guide uses tax year 2025 and IRS Publication 946 (2025) and the Instructions for Form 4562 (2025). Figures for other years differ. It is exam preparation, not tax advice.

DHCP and DNS: DORA, relay, records, lookups

Fri, 09 Oct 2026 00:00:00 +0000

From plug-in to web page: what each service does

DHCP gives a host its IP configuration (address, mask, default gateway, DNS servers) and DNS turns names into addresses. The decision rule for troubleshooting is short: if a host has no valid address, or an address starting with 169.254, the problem is DHCP; if it can ping by IP address but not by name, the problem is DNS. This guide follows one laptop from the moment it is plugged in until a page loads, because exams ask about each step in that order.

Digital forensics and chain of custody

Fri, 09 Oct 2026 00:00:00 +0000

Chain of custody is the documented record of who held a piece of evidence, when, where it was stored and why it changed hands, from the moment it was found until it is used or destroyed. NIST’s glossary defines it as a process that tracks the movement of evidence through its collection, safeguarding and analysis lifecycle. The exam rule: evidence you cannot account for is evidence you may not be able to use, so on any forensic scenario the correct answer protects integrity first (collect in order of volatility, copy rather than touch the original, hash, document every transfer) and analysis comes later.

DLP explained: Microsoft Purview policies

Fri, 09 Oct 2026 00:00:00 +0000

Data loss prevention (DLP) in Microsoft Purview stops sensitive content from leaving through the channels people use every day: email, SharePoint, OneDrive, Teams, devices and cloud apps. The decision rule for exams: a sensitivity label protects the item itself, a retention setting controls how long it lives, and a DLP policy controls what users may do with it and where it may go. If a scenario says “prevent”, “block” or “stop sharing”, DLP is the answer.

Docker basics: images, containers, Dockerfile

Fri, 09 Oct 2026 00:00:00 +0000

Docker packages an application and everything it needs into an image, and runs that image as an isolated process called a container. An image is the immutable template; a container is a running instance of it. Anything that must survive the container goes in a volume, and anything a browser or another host must reach has to be published with -p. Most Docker exam questions come down to one of those three sentences.

EIC due diligence for paid preparers

Fri, 09 Oct 2026 00:00:00 +0000

The earned income credit (EIC) is a refundable credit for workers with earned income below a limit, and it is one of the items the IRS requires paid preparers to check with extra care. The decision rule: a client’s right to the credit is decided by the rules in Publication 596, and the preparer’s own exposure is decided by the due diligence rules, which require a Form 8867, a documented interview, the computation, and retained records. This guide uses tax year 2025, with the IRS sources Publication 596 (2025) and the Instructions for Form 8867 (revised 11/2025). It is exam preparation, not tax advice.

Embeddings and vector search for AI exams

Fri, 09 Oct 2026 00:00:00 +0000

An embedding is a list of numbers, produced by a model, that represents the meaning of a piece of text, an image or other content, so that similar items end up close together. Vector search embeds the query with the same model and returns the stored items nearest to it. The exam decision rule: use vector search when the user’s words differ from the document’s words, use keyword search when exact tokens matter (part numbers, names, codes), and use hybrid search when both kinds of query occur. This guide goes layer by layer, from the numbers up to the index.

Entra Connect Sync vs Cloud Sync

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Entra Connect Sync and Microsoft Entra Cloud Sync both copy users, groups and contacts from on-premises Active Directory into Microsoft Entra ID. The decision rule: choose Cloud Sync when the scenario stresses several disconnected forests, little infrastructure to maintain or agents that fail over; choose Connect Sync when it needs something only the full sync engine offers, such as custom sync rules, attribute merging across domains, very large scale or pass-through authentication set up by the installer. Microsoft’s own decision guide calls Cloud Sync the strategic direction and says new features are built mainly there, so “which is the modern option” usually points at Cloud Sync, but a hard requirement still wins over the trend.

Entra External ID: B2B vs B2C explained

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Entra External ID is the umbrella for everything that involves people outside your organization. The decision rule: if your own workforce needs to share apps and resources with partner users, use B2B collaboration in your workforce tenant; if you are publishing an app to consumers or business customers, use an external tenant. Azure AD B2C, the older customer-identity product, is no longer sold to new customers, so new designs use External ID. Exam outlines test the guest and cross-tenant side far more than B2C.

Entra ID Protection and risk policies

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Entra ID Protection detects identity-based risk, scores it, and feeds the result to Conditional Access so that policy can respond automatically. It tracks two different things: sign-in risk (this particular sign-in may not be the real user) and user risk (this account may be compromised). The decision rule for exams: sign-in risk is answered with multifactor authentication, user risk with a secure password change or risk remediation, and both need Microsoft Entra ID P2 for full use.

EtherChannel and LACP: bundle links on a switch

Fri, 09 Oct 2026 00:00:00 +0000

What an EtherChannel is, and the mode rule that decides most questions

An EtherChannel (also called a port channel or link aggregation group) bundles two or more physical switch links between the same two devices into one logical link. You get more bandwidth and the bundle survives the loss of a member link. LACP (Link Aggregation Control Protocol, IEEE 802.3ad, now maintained as 802.1AX) is the open protocol that negotiates the bundle. PAgP is the Cisco-proprietary alternative, and “on” skips negotiation entirely.

Filing status practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free filing status practice questions with explained answers, written as short scenarios for US tax year 2025. Filing status is the category that decides which tax rates, standard deduction and credit limits a federal return uses, and there are five: single, married filing jointly, married filing separately, head of household and qualifying surviving spouse. The topic is tested in SEE Part 1, the individuals part of the IRS Special Enrollment Examination. These questions are for exam practice and general learning, not tax advice, and they follow IRS Publication 501 for 2025.

Filing status: the five statuses and their tests

Fri, 09 Oct 2026 00:00:00 +0000

Filing status is the category that decides which tax rates, standard deduction and credit limits a return uses. There are five: single, married filing jointly, married filing separately, head of household and qualifying surviving spouse. The decision rule: fix the taxpayer’s marital status on the last day of the year first, then check the special statuses (surviving spouse, then head of household) before falling back to single or married filing separately. This guide uses tax year 2025 and IRS Publication 501 (2025). Rules and amounts can change by year, so a different exam year needs a fresh check.

Fine-tuning vs RAG vs prompting

Fri, 09 Oct 2026 00:00:00 +0000

Prompting, retrieval-augmented generation (RAG) and fine-tuning are three ways to make a general model fit a specific job, and they fix three different problems. Prompting changes the instructions for one request. RAG changes what the model can see by fetching your content at question time. Fine-tuning changes the model itself by training it on examples. The decision rule: if the information changes or must come from your documents, use RAG; if the behaviour must change (tone, format, a narrow skill), consider fine-tuning; if neither is true, start with a better prompt, because it is the cheapest option and the first one to try.

Git branching strategies compared

Fri, 09 Oct 2026 00:00:00 +0000

A branching strategy is the team’s agreement on where work happens, how it reaches the main line and how releases are cut. The decision rule is simple: the more often you release one version, the shorter your branches should be; the more versions you must support at once, the more release branches you need. Everything else is detail on those two axes.

The four models side by side

Model	Branches	Merge path	Release cadence it suits
Trunk-based	One shared trunk, plus very short-lived branches	Into trunk, at least daily, unfinished work behind feature flags	Many deployments per day, strong automated tests
GitHub flow	`main` plus one branch per change	Pull request, review, merge, delete branch	Continuous delivery of one live version
GitFlow	`main`, `develop`, `feature/`, `release/`, `hotfix/*`	Features into `develop`; releases and hotfixes into both `main` and `develop`	Scheduled, versioned releases
Release branches (as Microsoft describes them)	Feature branches plus a long-lived branch per supported release	Fixes land in the release branch and are cherry-picked back to `main`	Products with several shipped versions

What each model actually does

Trunk-based development. Developers share a single branch, the trunk, and avoid long-lived development branches. The trunkbaseddevelopment.com site says team members should commit to the trunk at least once every 24 hours. Very small teams may commit straight to it; larger teams use short-lived branches with pull requests and a build server that catches breakage. Because half-finished work lands in the trunk, feature flags hide it until it is ready. Releases either come straight from the trunk with fixes rolled forward, or from a release branch cut just in time and deleted afterwards.

HSRP and VRRP: first hop redundancy explained

Fri, 09 Oct 2026 00:00:00 +0000

The problem and the answer in one paragraph

Hosts have one default gateway address and cannot choose another when it dies. A First Hop Redundancy Protocol (FHRP) lets two or more routers share one virtual IP address (and a virtual MAC address), so the hosts keep pointing at the same gateway while the routers decide among themselves which one answers. HSRP (Hot Standby Router Protocol) is Cisco-proprietary and VRRP (Virtual Router Redundancy Protocol, RFC 5798) is the open standard. The decision rule: highest priority wins the active (HSRP) or master (VRRP) role, and a higher-priority router only takes the role back after a failure if preemption is on. Preemption is off by default in HSRP and on by default in VRRP, which is the detail exam questions love.

IaC: declarative vs imperative explained

Fri, 09 Oct 2026 00:00:00 +0000

Infrastructure as code (IaC) means describing infrastructure in text files kept in version control, so a change is a commit instead of a series of clicks. A declarative definition states the end result and lets the tool work out the steps; an imperative script lists the steps and leaves you to handle the starting state. If an exam question says “desired state”, “repeatable” or “the same result no matter how many times it runs”, the answer is declarative and idempotent. Bicep, ARM templates and Terraform are declarative. A Bash or PowerShell script that calls az commands in order is imperative.

Incident response lifecycle: order of steps

Fri, 09 Oct 2026 00:00:00 +0000

The incident response lifecycle is the ordered set of phases a team follows from getting ready for an incident to learning from it. In the classic NIST model (SP 800-61 Rev. 2) the phases are preparation, detection and analysis, containment, eradication and recovery, and post-incident activity. The decision rule for the exam: find the phase the scenario is in, then choose the action that belongs to the next one. Contain before you eradicate, eradicate before you recover, and never skip the review at the end.

Insider Risk Management in Purview

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Purview Insider Risk Management (IRM) correlates activity signals from Microsoft 365 to find potentially malicious or inadvertent risky behavior by users, such as data leaks, IP theft and security violations, and lets analysts triage alerts, open cases and act. The decision rule: DLP blocks or warns on a single action against sensitive content, while IRM looks at patterns of behavior over time and asks whether a person is becoming a risk. Microsoft builds it privacy-first: users are pseudonymized by default, with role-based access and audit logs on the analysts.

IPv4 vs IPv6: addressing, SLAAC, dual stack

Fri, 09 Oct 2026 00:00:00 +0000

The short answer, then the mechanics

IPv4 uses 32-bit addresses, IPv6 uses 128-bit addresses, and nearly everything else that exams ask follows from that and from three design changes: IPv6 has no broadcast (multicast replaces it), hosts can configure themselves with SLAAC (stateless address autoconfiguration) from router advertisements, and every interface carries a link-local address (fe80::/10) that is used for neighbor discovery and routing next hops. The decision rule for scenario questions: for “a host must get an address with no DHCP server” think SLAAC, for “which address is valid only on this link” think link-local, and for “run both protocols side by side during migration” think dual stack. The addressing architecture is RFC 4291, and the address length, link-local, multicast and no-broadcast facts below come from it.

JSON vs YAML vs XML: formats and pitfalls

Fri, 09 Oct 2026 00:00:00 +0000

Which format, in one rule

JSON is the default format of REST APIs; YAML is what people write by hand for configuration (Ansible playbooks, Kubernetes manifests, pipeline files); XML is the older, more verbose format that survives in NETCONF, SAML and SOAP. They all describe the same kinds of structure, nested key-value data and lists, so exam questions test whether you can read one, convert it to another, and spot what breaks it. The rule: machines talking to machines use JSON, humans editing files use YAML, and protocols standardised before about 2010 use XML.

KQL cheat sheet for certification exams

Fri, 09 Oct 2026 00:00:00 +0000

Kusto Query Language (KQL) is the read-only query language of Azure Monitor Logs, Microsoft Sentinel, Defender XDR advanced hunting, Microsoft Fabric Eventhouses and Azure Data Explorer. A query starts with a table and passes its rows through operators separated by a pipe (|), and each operator works on the output of the one before it. The decision rule for exam questions: filter first, aggregate second, shape the output last, and prefer the operator that matches whole terms (has) over the one that scans substrings (contains).

KQL practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free KQL practice questions with explained answers, written as short scenarios on filtering, aggregating, joining and parsing log data. Kusto Query Language is the read-only query language of Azure Monitor Logs, Microsoft Sentinel, Defender XDR advanced hunting and Fabric Eventhouses, and each question states the table columns it uses. The topic is tested in SC-200, DP-600 and DP-700, and in a lighter form in AZ-104.

How to use these questions

Read the schema in the stem first, then trace the query one operator at a time and write down which columns exist after each pipe. Commit to one letter before opening the explanation. Underline the constraint word, such as “whole term”, “never”, “latest” or “per user”, because it usually picks the operator. The guide KQL cheat sheet for certification exams summarises the operators used here.

Kubernetes and AKS basics for exams

Fri, 09 Oct 2026 00:00:00 +0000

Kubernetes is an open-source orchestrator that keeps containers running across a group of machines, in the shape you declare. Azure Kubernetes Service (AKS) is the managed version: Azure runs the control plane, you run the workloads and, in AKS Standard, the node pools. On an exam, choose AKS when the scenario needs the full Kubernetes API, existing manifests or Helm charts, or fine control of scheduling and networking; choose Container Apps when the team wants containers and scaling without administering Kubernetes. If you do not know what a container or an image is yet, read Docker basics first.

Lakeflow Declarative Pipelines (ex-DLT)

Fri, 09 Oct 2026 00:00:00 +0000

Lakeflow Declarative Pipelines is the Databricks framework, formerly called Delta Live Tables (DLT), in which you declare the tables you want and the platform works out the order, retries and incremental processing. The decision rule for exams: if a scenario is a chain of table-to-table transformations with data quality rules and the team does not want to hand-write orchestration or checkpoint handling, the answer is a declarative pipeline. If the work is arbitrary code, external calls or steps that do not produce tables, the answer is a notebook or task in Lakeflow Jobs.

Lakehouse vs warehouse: Fabric and Databricks

Fri, 09 Oct 2026 00:00:00 +0000

A lakehouse keeps files and Delta tables in open storage and is built mainly with Spark; a warehouse is a relational store built mainly with T-SQL. In Microsoft Fabric both keep their tables as Delta in OneLake and share the same SQL engine, so the choice is about how the team works, not where the data sits. The decision rule: Spark or unstructured data means lakehouse, T-SQL with multi-table transactions means warehouse, and when the scenario is on Databricks, the answer is almost always the Databricks lakehouse.

LLM Evaluation: Metrics and LLM-as-Judge

Fri, 09 Oct 2026 00:00:00 +0000

LLM evaluation measures whether a generative application is good enough to ship and stays good once running. The decision rule is: use reference-based metrics (ROUGE, BLEU, BERTScore) when a known correct answer exists, use reference-free judged metrics (groundedness, relevance, coherence) when it does not, and always check retrieval separately from generation. Accuracy, the classifier’s metric, rarely applies, because there is usually no single right answer.

Two moments: offline and online

Offline evaluation runs before release on a fixed test set: questions, optionally reference answers and retrieved context, run against each version of the prompt, model or index. The set is fixed so a score change can be attributed to the change you made. Online evaluation scores a sample of real production traffic on a schedule, plus user feedback and operational data such as latency and token cost. Offline answers “should this version ship?”, online answers “is it still good?” and catches things a test set missed, such as new question types or a stale document index. A mature setup uses both, and offline evaluation runs automatically in a CI pipeline with thresholds that fail the build (see CI/CD explained).

Managed identities vs service principals

Fri, 09 Oct 2026 00:00:00 +0000

A service principal is the identity an application has inside a Microsoft Entra tenant; a managed identity is a special kind of service principal whose credentials Azure creates, holds and rotates for you. The decision rule for exams: if the code runs on an Azure resource that supports managed identities, use a managed identity; if it runs outside Azure and has to authenticate with no stored secret, use workload identity federation; only when neither is possible use a service principal with a certificate or secret.

Medallion architecture: bronze, silver, gold

Fri, 09 Oct 2026 00:00:00 +0000

Medallion architecture is a pattern for organising lakehouse tables into three layers of rising quality: bronze holds raw data as it arrived, silver holds validated and cleaned records, and gold holds business-ready, usually aggregated tables. The decision rule for exams: raw and replayable means bronze, cleaned one-row-per-record means silver, shaped for a report or dashboard means gold. Nothing downstream writes back into an earlier layer.

One order, followed through three layers

This guide uses a single small example to show what each layer does. A retailer receives order events as JSON files, sometimes twice, sometimes with a missing customer, sometimes with test orders.

MFA, passwordless and passkeys explained

Fri, 09 Oct 2026 00:00:00 +0000

Multifactor authentication (MFA) means proving identity with two or more different kinds of evidence (something you know, have or are). Passwordless means the password is removed from the sign-in altogether. A passkey is a passwordless credential built on FIDO standards that cannot be phished, because it only answers to the website it was created for. The decision rule for exams: when a question asks for the strongest option or says “phishing-resistant”, pick a passkey (FIDO2 security key, Windows Hello for Business, or a passkey in Microsoft Authenticator); when it asks only for “MFA”, almost any second factor qualifies.

Microsoft Sentinel explained for exams

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Sentinel is Microsoft’s cloud-native SIEM (it collects security data, detects threats and groups alerts into incidents) with SOAR abilities (automation rules and playbooks respond to them). The decision rule: if a question says collect and correlate logs from many sources, detect with queries, investigate incidents and automate the response, the answer is Sentinel; if it says protect one workload and give recommendations on its configuration, the answer is a Defender product. Sentinel stores its data in a Log Analytics workspace and you query it with KQL.

MLflow basics: tracking, models, registry

Fri, 09 Oct 2026 00:00:00 +0000

MLflow is an open-source framework that records how a machine learning model was built (tracking), packages it in a standard format (models) and manages its versions (registry). The decision rule for exams: if a scenario says compare training attempts, reproduce a result, or know which version is deployed, the answer is MLflow: tracking for experiments, the registry for lifecycle. The same MLflow API is used on Databricks, Amazon SageMaker AI and Azure Machine Learning, which is why it appears in exams from all three vendors.

MLOps Basics: Registry, CI/CD, Drift

Fri, 09 Oct 2026 00:00:00 +0000

MLOps applies DevOps discipline to machine learning: every model is versioned in a registry, built and promoted by an automated pipeline, and monitored in production, because a model degrades even when no code changes. The decision rule for exam scenarios is: if quality falls while the infrastructure is healthy and nothing was deployed, the answer is drift monitoring plus a retraining path; if a change must reach production safely, the answer is a pipeline with a validation gate and a registry, not a manual copy.

NAT practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free NAT practice questions with explained answers, written for anyone preparing for the Cisco CCNA. Network Address Translation (NAT) rewrites IP addresses as packets cross a router, so private inside hosts can reach public networks; the CCNA tests inside source NAT with static entries, pools and PAT (overload). The topic appears in the IP Services domain of CCNA 200-301. All questions are original and use Cisco IOS behavior.

How to use these questions

Answer each question before opening the explanation, and commit to one letter. Underline the constraint word in the stem, such as “many hosts”, “permanent” or “without overload”, because it usually picks the NAT type. Read the explanation even when the answer was right, since each one also says why the other three options fail. The concepts behind the questions are covered in the NAT types topic guide.

NAT types: static, dynamic and PAT

Fri, 09 Oct 2026 00:00:00 +0000

Which NAT type fits the requirement

Network Address Translation (NAT, RFC 3022 describes the traditional form) rewrites IP addresses as packets cross a router, mainly so hosts with private addresses can reach the internet. Three types matter. Static NAT is a permanent one-to-one mapping, used when something outside must reach a specific inside host. Dynamic NAT maps inside hosts to addresses taken from a pool as they need them. PAT (port address translation, the overload keyword; RFC 3022 calls it NAPT) lets many inside hosts share one public address by also rewriting the source port. The decision rule: one server reachable from outside means static, many users and a short supply of public addresses means PAT, and dynamic with a pool is the in-between that needs one public address per simultaneous host.

Network port practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free network port practice questions with explained answers, checked against the IANA service name and port number registry. A port number identifies the service a TCP or UDP packet is addressed to, and exams expect the well-known ports plus the secure alternative for each insecure protocol. The topic is tested in CompTIA Network+ (N10-009) and CompTIA Security+ (SY0-801).

How to use these questions

Answer first, then open the explanation. Note the constraint in the stem: transport (TCP or UDP), direction (client to server or server to client), and whether encryption is required. Wrong options are explained too, so they double as a mini port list. The common ports cheat sheet holds the full table for review.

Network troubleshooting: commands and method

Fri, 09 Oct 2026 00:00:00 +0000

Method first, then the command that answers the question

Network troubleshooting questions test two things: the order of the steps and which tool answers which question. The method, in the order CompTIA’s N10-009 outline lists it: identify the problem, establish a theory of probable cause, test the theory, establish a plan of action (including its side effects), implement the fix or escalate, verify full functionality and put preventive measures in place, then document what happened. The decision rule for any “what should the technician do next” question: take the step that follows the one already completed, and never jump to the fix before a theory has been tested.

NIST CSF, SP 800-53, ISO 27001, CIS Controls

Fri, 09 Oct 2026 00:00:00 +0000

Four documents come up whenever a security exam says “framework”: the NIST Cybersecurity Framework (CSF) 2.0, NIST SP 800-53, ISO/IEC 27001 and the CIS Controls. They are not interchangeable. CSF 2.0 organises cybersecurity outcomes and is used to plan and communicate a program. SP 800-53 is a detailed catalog of controls. ISO/IEC 27001 is a management-system standard against which an organisation can be audited and certified. The CIS Controls are a short, prioritised list of practical safeguards. The decision rule: match the verb in the question. Organise or communicate points to CSF, select detailed controls to SP 800-53, get certified to ISO 27001, and implement the most important safeguards first to CIS.

NSG and VNet practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free NSG and virtual network practice questions with explained answers, built as short packet-tracing and subnet-sizing scenarios. A network security group is an ordered list of allow and deny rules that Azure applies to traffic in a virtual network, and a subnet is the address range inside the network where resources receive their IPs. The topic is tested in AZ-104, SC-500 and, at concept level, SC-900.

How to use these questions

Attempt each question before opening the explanation. For rule questions, sort the rules by priority number first and walk the packet through them in that order, as Azure does. For sizing questions, calculate the total with 2 to the power of (32 minus the prefix) and subtract the five addresses Azure reserves. Note the constraint word in the stem (inbound or outbound, new or existing, least). The guide NSG rule evaluation in Azure covers the theory.

NSG rule evaluation in Azure

Fri, 09 Oct 2026 00:00:00 +0000

A network security group (NSG) is a list of allow and deny rules; Azure checks a packet against the rules in priority order, lowest number first, and stops at the first match. If nothing you wrote matches, three default rules decide: traffic inside the virtual network is allowed, Azure load balancer probes are allowed, and everything else inbound is denied. The rest of this page traces packets through that logic.

OneLake shortcuts: rules, limits, security

Fri, 09 Oct 2026 00:00:00 +0000

A OneLake shortcut is a pointer, shown as a folder, that makes data stored somewhere else appear inside a Fabric lakehouse or KQL database without copying it. The decision rule for exam questions: if the scenario says “without copying”, “no ingestion pipeline” or “keep the data where it is”, the answer is a shortcut; if it says the source is an operational database that must be replicated, the answer is mirroring; if it says shortcut queries are too slow in an Eventhouse, the answer is query acceleration on the shortcut.

OSI model practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free OSI model practice questions with explained answers, written as short scenarios rather than memory drills. The OSI model splits network communication into seven layers, and the exam skill is placing a symptom, a device or a protocol data unit on the correct layer. The material is tested in CompTIA Network+ (N10-009), where objective 1.1 covers the layers directly.

How to use these questions

Answer each question before opening the explanation, and commit to one letter. Underline the constraint word in the stem (for example “first”, “only”, “most likely”), because it usually separates the right option from a plausible one. Wrong options are explained as well, so read the whole explanation even after a correct answer. The OSI vs TCP/IP guide covers the theory behind every question.

OSI vs TCP/IP model: layers and mapping

Fri, 09 Oct 2026 00:00:00 +0000

Seven layers or four, and which one the exam wants

The OSI model has seven layers and is the reference model that exams use to talk about where something happens. The TCP/IP model has four layers and describes how the protocols that run the internet are actually organised. They describe the same traffic, so the rule is: use the OSI numbers when a question says “Layer 2”, “Layer 3” or “Layer 4”, and use TCP/IP only when a question asks about the protocol suite itself. The OSI model is the one you must be able to recite and apply; the TCP/IP model is the one that maps to real packets.

OSPF explained: states, DR/BDR and cost

Fri, 09 Oct 2026 00:00:00 +0000

OSPF in one paragraph

OSPF (Open Shortest Path First) is a link-state routing protocol: every router in an area builds the same map of the network, called the link-state database, and each router runs the shortest path algorithm on it to choose routes by lowest total cost. Two routers exchange routes only after they become neighbors and then fully adjacent, so nearly every OSPF exam question is really about what has to match for that to happen and where the process stalls. The decision rule: if the neighbor is missing or stuck, compare the settings that must match; if the route is wrong, add up the interface costs.

OSPF practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free OSPF practice questions with explained answers, covering neighbor states, what must match between routers, DR and BDR election, router ID selection and cost. OSPF (Open Shortest Path First) is a link-state routing protocol in which routers share a map of the network and choose the lowest-cost path. It is tested on the CCNA 200-301 as single-area OSPFv2 and named in lighter form on CompTIA Network+ N10-009. All questions assume OSPFv2 on Cisco IOS with default settings unless the stem says otherwise.

Overfitting and ML Metrics Explained

Fri, 09 Oct 2026 00:00:00 +0000

Overfitting means a model has memorised its training data and scores far better there than on data it has not seen; underfitting means it is too simple to score well anywhere. The decision rule for the metric questions is: pick the metric by the cost of the mistake. If a missed positive is expensive, optimise recall; if a false alarm is expensive, optimise precision; if the classes are imbalanced, never trust accuracy alone.

Performance-based question practice tasks

Fri, 09 Oct 2026 00:00:00 +0000

A performance-based question (PBQ) asks you to do a small task in a simulated interface instead of picking a lettered option: order rules, read a log, drag items into place or fill in a setting. The decision rule is to treat each one as a short scenario with a checkable result: work out the answer on paper first, then make the interface match it. The four tasks below are original exercises written for practice, not copies or reconstructions of any real exam item, and each one trains a skill that PBQs commonly test.

Power Query basics: steps, M and query folding

Fri, 09 Oct 2026 00:00:00 +0000

Power Query is the data preparation engine behind Power BI Desktop, Excel, dataflows and Fabric Dataflow Gen2. You click transformations, it records each one as a step, and the steps are written in a formula language called M. The decision rule for exam questions: merge adds columns, append adds rows, and a transformation that can be translated into the source’s own query language “folds” and runs at the source, which is faster than running it in Power Query. Most scenario questions are one of those three ideas in disguise.

Preparer penalties: IRC 6694 and 6695

Fri, 09 Oct 2026 00:00:00 +0000

The Internal Revenue Code penalises paid return preparers in two main ways. Section 6694 punishes a return or refund claim with an understatement caused by a bad position, and the amount depends on how bad: negligent or unreasonable under §6694(a), willful or reckless under §6694(b). Section 6695 punishes failures of procedure, such as not signing, not giving the client a copy, or not meeting due diligence. The decision rule: if the error is in the tax answer, think §6694; if it is in the paperwork or the checking, think §6695.

Private endpoint vs service endpoint in Azure

Fri, 09 Oct 2026 00:00:00 +0000

A private endpoint puts a network interface with a private IP address from your subnet in front of an Azure service, so the service gets an address inside your network. A service endpoint keeps the service on its public address but tells it which subnet the traffic comes from, so the service can accept only that subnet. The decision rule: if the scenario needs a private IP, access from on-premises, or public access fully disabled, pick the private endpoint. If it needs a cheap, no-DNS restriction to certain VNet subnets, a service endpoint works. Microsoft itself recommends Private Link and private endpoints for secure and private access to platform services.

Privileged Identity Management (PIM)

Fri, 09 Oct 2026 00:00:00 +0000

Privileged Identity Management (PIM) in Microsoft Entra turns permanent administrator access into just-in-time access: a person is made eligible for a role, and has to activate it (with MFA, a justification, possibly approval) for a limited time before using it. The decision rule for exams: if the requirement is “only when needed”, “time-limited”, “with approval” or “with an audit trail” for a privileged role, the answer is PIM. If it is about the conditions of a sign-in, it is Conditional Access.

Prompt engineering basics: five fixes in order

Fri, 09 Oct 2026 00:00:00 +0000

Prompt engineering is the practice of shaping the input to a language model so that its output is useful, correctly formatted and based on the right information, without changing the model itself. The exam decision rule: fix the prompt before anything else. Give clear instructions, add examples if the format is wrong, supply source material if facts are wrong, and lower the temperature if the output must be repeatable. Only when none of that works do exams move on to RAG or fine-tuning.

Purview eDiscovery: search, hold, review, export

Fri, 09 Oct 2026 00:00:00 +0000

Microsoft Purview eDiscovery finds, preserves, reviews and exports content from Exchange Online, Teams, Microsoft 365 Groups, OneDrive, SharePoint and Viva Engage so it can serve as evidence in an investigation or legal matter. The decision rule: to find content, search; to stop it from being deleted while you work, put a hold on the case’s data sources; to tag, analyze and cull it, you need Premium features and a review set. Audit answers “who did what”, eDiscovery answers “show me the items”.

Python for network engineers: the exam subset

Fri, 09 Oct 2026 00:00:00 +0000

Python on the 200-901 CCNA Automation exam (formerly DevNet Associate) is not a programming test. It is a reading test: you get a short script or a code fragment and must say what it prints, what it sends, or why it fails. The decision rule is to know seven things cold: the four core data types (list, tuple, dict, string), how to index them, loops and comprehensions, defining and calling a function, json.loads against json.dumps, one requests call with its status handling, and which except clause catches which error. Everything below is that subset, and every snippet in this guide was run before it was written down.

Qualitative vs quantitative risk: SLE, ARO, ALE

Fri, 09 Oct 2026 00:00:00 +0000

Qualitative risk analysis rates risks with words or scores such as low, medium and high, based on judgement of likelihood and impact. Quantitative risk analysis puts numbers and money on them. The standard quantitative formulas are SLE = asset value x exposure factor, and ALE = SLE x ARO. SLE is the single loss expectancy, ARO is the annualised rate of occurrence and ALE is the annualised loss expectancy. The decision rule: use qualitative when you need a fast ranking or lack data, use quantitative when someone needs a dollar figure to justify spending, and always compare the yearly risk reduction with the yearly cost of the control.

RAG explained: chunking, embeddings, retrieval

Fri, 09 Oct 2026 00:00:00 +0000

Retrieval-augmented generation (RAG) is a pattern in which an application first searches your own content for passages relevant to a question, then gives those passages to a large language model together with the question, so the model answers from your material instead of from memory. The exam decision rule: if the information changes or must be grounded in your documents and cited, use RAG; if the model’s behaviour or style must change, that is fine-tuning. RAG needs no retraining when the content changes, you only update the index.

RAG practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free RAG practice questions with explained answers, written for anyone preparing for a Microsoft AI exam. Retrieval-augmented generation (RAG) searches your own content for passages that match a question and hands them to a large language model together with the question, so the answer is grounded in your documents instead of the model’s memory. The pattern is tested in the generative AI skill area of AI-103. All questions are original and use Azure AI Search and Microsoft Foundry terminology.

Responsible AI principles: Microsoft vs AWS

Fri, 09 Oct 2026 00:00:00 +0000

Responsible AI is the set of principles and controls that make an AI system fair, safe, understandable and accountable. Microsoft’s Responsible AI Standard rests on six principles: fairness, reliability and safety, privacy and security, inclusiveness, transparency and accountability. AWS describes eight dimensions: fairness, explainability, privacy and security, safety, controllability, veracity and robustness, governance and transparency. The exam decision rule: read the scenario for the one harm it describes, then name the principle that harm belongs to, and when the fix is asked for, prefer the human or technical control that targets that harm.

REST API basics: methods, status codes, auth

Fri, 09 Oct 2026 00:00:00 +0000

A REST call is a method, a URL and a status code to read

A REST API exposes resources (devices, users, VLANs) at URLs and lets a client act on them with standard HTTP methods; the server answers with a status code, headers and usually a JSON body. The rule that settles most questions: the method says what you want done, the URL says to what, and the status code family says who is at fault, with 2xx meaning success, 4xx meaning the client’s request is wrong, and 5xx meaning the server failed. HTTP semantics are defined in RFC 9110; REST itself is a style on top, not a standard.

REST vs NETCONF vs RESTCONF and YANG

Fri, 09 Oct 2026 00:00:00 +0000

The rule: YANG is the model, NETCONF and RESTCONF are two ways to reach it

A plain REST API is whatever a vendor decided to publish: its own URLs, its own JSON fields, usually on a controller. NETCONF (RFC 6241) and RESTCONF (RFC 8040) are standard protocols that configure a device from a YANG data model instead of CLI text. NETCONF speaks XML remote procedure calls over SSH; RESTCONF maps the same YANG data onto HTTPS methods with JSON or XML bodies. The decision rule for exam questions: if the question says XML over SSH, port 830, or candidate datastore and commit, the answer is NETCONF; if it says HTTP methods, URLs and JSON against a YANG-modelled device, it is RESTCONF; if it says a controller or cloud service with its own documented endpoints, it is a REST API.

Retention labels vs retention policies

Fri, 09 Oct 2026 00:00:00 +0000

A Microsoft Purview retention policy applies a keep-or-delete rule to whole locations (all mailboxes, all SharePoint sites, a set of Teams chats), while a retention label applies a rule to one individual item and travels with it. The decision rule: if the rule is the same for everything in a location, use a policy; if different items in the same place need different treatment, or the item must be a record, use a label. Both can apply to the same item, and when they overlap, retention beats deletion and the longest retention period wins.

Risk calculation practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free risk calculation practice questions with explained answers, covering single loss expectancy (SLE), annualized rate of occurrence (ARO), annualized loss expectancy (ALE), cost-of-control decisions and the four risk treatments. Quantitative risk analysis puts a money value on a risk using SLE = asset value x exposure factor and ALE = SLE x ARO. The calculations appear in objective 5.2 of Security+ SY0-801, and every figure below was recomputed in code.

Row-level security: Power BI, Fabric and SQL

Fri, 09 Oct 2026 00:00:00 +0000

Row-level security (RLS) makes the same table return different rows to different people, based on who is asking. The decision rule for exams: rows are hidden by RLS, columns and tables by object-level or column-level security, and the product decides the mechanism. In a Power BI semantic model RLS is a role with a DAX filter, in a Fabric warehouse or SQL database it is a security policy with a predicate function. Use static rules for a handful of fixed groups and dynamic rules, driven by the signed-in user, when each person sees something different.

SageMaker AI components along the ML lifecycle

Fri, 09 Oct 2026 00:00:00 +0000

Amazon SageMaker AI is a set of managed components, one or more per stage of the machine learning lifecycle: prepare data, train and tune, track and register, deploy, monitor. The decision rule: read the scenario’s lifecycle stage first, then pick the component built for that stage. Feature Store stores features, Pipelines automates steps, Model Registry versions and approves models, and endpoints or batch transform serve predictions. Since 2026, several well-known components are also closed to new customers, which matters for questions written before that.

SAS tokens and stored access policies

Fri, 09 Oct 2026 00:00:00 +0000

A shared access signature (SAS) is a signed URL token that gives a client limited access to storage resources: which resource, which permissions, for how long. The token is not stored by Azure Storage; the service only checks the signature. The decision rule for exams: the key that signed the token decides how you can revoke it. An ad hoc service or account SAS dies only when its signing account key is rotated or it expires; a service SAS tied to a stored access policy is revoked by changing or deleting the policy; a user delegation SAS is revoked by revoking the user delegation key or the signer’s role assignments.

Schedule C and self-employment tax

Fri, 09 Oct 2026 00:00:00 +0000

Schedule C (Form 1040) reports the profit or loss of a sole proprietorship, and Schedule SE turns that profit into self-employment (SE) tax. The decision rule: net profit goes to the 1040 as income, 92.35% of it is the base for SE tax at 15.3% up to the Social Security wage base, and half of the SE tax comes back as an adjustment to income. This guide uses tax year 2025 and the IRS sources Publication 334 (2025), the Instructions for Schedule C and the Instructions for Schedule SE. It is exam preparation, not tax advice.

Security control categories and types

Fri, 09 Oct 2026 00:00:00 +0000

Two questions, two answers: category and type

Every security control can be described on two independent axes. The category says what the control is made of and who or what operates it: technical, managerial, physical or operational. The type says what the control does relative to an incident: preventive, deterring, detective, corrective, compensating or directive. The decision rule for exams: read which axis the question asks about, then classify on that axis only. A security camera is physical in category and detective in type, and both statements are true at once.

Security controls practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free security controls practice questions with explained answers. Security controls are classified in two ways: by category (technical, managerial, operational or physical) and by type (preventive, deterring, detective, corrective, compensating or directive). The same control can be described on both axes, so each stem states which one it asks for. Both appear in objective 1.1 of Security+ SY0-801.

How to use these questions

Attempt each question before opening the answer, then read why each wrong option fails.
Note whether the stem asks for the category, the type, or both.
Ask what the control does at the moment of an incident: stops it, discourages it, spots it, repairs it, or stands in for a missing control.
The topic guide security control categories explains both axes with examples.

Warm-up: categories and basic types

Question 1. A firewall rule drops inbound traffic to the database port from every network except the application servers. In which control category does this fall?

Security+ acronyms grouped by exam domain

Fri, 09 Oct 2026 00:00:00 +0000

Learn them by domain, not alphabetically

Security+ questions reuse about a hundred acronyms, and the exam rarely spells them out. The practical way to learn them is to attach each one to the domain where it is used, because the domain tells you the context the question is set in. This guide groups the acronyms that appear in the SY0-801 objectives by the five domains. The decision rule: when two acronyms look alike, work out which stage of the work each belongs to (finding, scoring, fixing, reporting) before reading the options.

Semantic model storage modes: Import, DirectQuery

Fri, 09 Oct 2026 00:00:00 +0000

A storage mode is a property of each table in a Power BI or Fabric semantic model. It decides whether the data is copied into the model’s memory (Import), left in the source and queried live (DirectQuery), read from Delta tables in OneLake on demand (Direct Lake), or either of the first two depending on the query (Dual). The decision rule for exams: pick the mode from the constraint in the scenario. “Fastest visuals, data can be hours old” is Import. “Must be current, or too big to copy, source is a database” is DirectQuery. “Large data already in a Fabric lakehouse or warehouse, no long refresh” is Direct Lake.

Sensitivity labels and DLP practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Microsoft Purview practice questions with explained answers, written as short scenarios about sensitivity labels, label policies, auto-labeling and data loss prevention (DLP). A sensitivity label classifies and protects the item itself, while a DLP policy controls what users may do with sensitive items and where they may send them. The topic is tested in SC-401, SC-900 and AB-900, and parts of it appear in the Power BI and Fabric exams.

How to use these questions

Try each question before opening the explanation and commit to one letter. Underline the constraint word in the stem, such as “published”, “recommend”, “simulation”, “already applied” or “block”, because it usually points to the one layer that is failing: the label, its scope, the policy, the application method or the DLP rule. Read why each wrong option fails, since the same distractors return on real exams. The guide sensitivity labels explained holds the theory behind these fifteen.

Sensitivity labels explained

Fri, 09 Oct 2026 00:00:00 +0000

A Microsoft Purview sensitivity label is a tag stored in the metadata of a file or email that classifies it (for example Public, General, Confidential) and can enforce protection such as encryption and content markings that stay with the content wherever it goes. The decision rule: the label describes and protects the item itself, DLP controls where the item may go, and retention controls how long it lives. A label does nothing for users until it is published in a label policy.

Shared Responsibility Model: IaaS, PaaS, SaaS

Fri, 09 Oct 2026 00:00:00 +0000

Who owns each layer: the rule and the matrix

In the shared responsibility model the provider secures the cloud infrastructure and you secure what you put in it, and the further you move from IaaS toward SaaS, the more layers the provider takes over, but you always keep your data, your identities and your configuration. The decision rule for a question is: find the layer the scenario is about, find the service model, and read the answer off the matrix below. If the layer is data, accounts or access, the answer is the customer in every model.

SIEM vs SOAR vs XDR: which tool does what

Fri, 09 Oct 2026 00:00:00 +0000

A SIEM (security information and event management) collects logs from many sources, correlates them and raises alerts. A SOAR (security orchestration, automation and response) runs automated workflows, called playbooks, that act on those alerts. An XDR (extended detection and response) joins detections from one vendor’s endpoint, identity, email and cloud products into single incidents and responds on those same products. The decision rule: if the question is about collecting and correlating data from anywhere, the answer is SIEM; if it is about automating a repeatable response, SOAR; if it is about one incident spanning the protected product set with built-in response, XDR.

Spanning Tree practice questions with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Spanning Tree practice questions with explained answers, written for CCNA candidates. Spanning Tree Protocol (STP) prevents Layer 2 loops in switched networks with redundant links by electing a root bridge and blocking the ports that would create a second path; Cisco switches typically run the rapid version, Rapid PVST+. The topic is tested in the Network Access domain of CCNA 200-301. All questions are original and use Cisco IOS behavior.

Spanning Tree Protocol: roles, states, root

Fri, 09 Oct 2026 00:00:00 +0000

What spanning tree does, and the order it decides in

Spanning Tree Protocol (STP, IEEE 802.1D, with the rapid version 802.1w) keeps a switched network with redundant links free of Layer 2 loops by electing one root bridge and blocking every port that would create a second path to it. Switches exchange BPDUs (bridge protocol data units) to do this, and the result is always decided in the same order: root bridge first, then one root port per non-root switch, then one designated port per segment, and everything left over blocks. The decision rule for any exam question: lowest bridge ID wins the root, lowest path cost to the root wins the root port, and ties are broken by lowest sender bridge ID, then lowest sender port ID.

Spark basics: transformations, actions, shuffles

Fri, 09 Oct 2026 00:00:00 +0000

Apache Spark is an engine that splits a dataset into partitions and processes them in parallel across a cluster. The decision rule for exams: transformations are lazy and only describe the work; an action triggers execution; and any operation that must regroup rows across partitions (a join, a group-by, a repartition) causes a shuffle, which is the expensive part. Nearly every Spark question, from the Databricks Spark Developer exam to Fabric and Databricks data engineering, comes down to spotting which of those three things is happening.

SQL joins and window functions with results

Fri, 09 Oct 2026 00:00:00 +0000

A join combines columns from two tables by matching rows, and a window function calculates across a set of related rows without collapsing them into one. The decision rule: use a join when you need columns from another table, use a window function when you need a per-group calculation (rank, running total, previous value) and still want every row back. GROUP BY collapses rows, a window function with OVER does not.

SSO, SAML, OAuth and OIDC explained

Fri, 09 Oct 2026 00:00:00 +0000

Single sign-on (SSO) is the outcome: sign in once, reach many applications. SAML and OpenID Connect (OIDC) are the protocols that deliver it by proving who the user is. OAuth 2.0 is different: it grants an application limited access to something, such as an API, and says nothing on its own about who signed in. The decision rule for the exam: “who is this user” means SAML or OIDC; “let this app act on the user’s data” means OAuth; and OIDC is OAuth with an identity layer added on top.

Standard vs itemized deduction: how to choose

Fri, 09 Oct 2026 00:00:00 +0000

Every individual return subtracts either the standard deduction or itemized deductions from adjusted gross income, never both. The decision rule: compute both and use the larger, unless the taxpayer is not allowed the standard deduction at all, in which case itemizing is the only option. This guide uses tax year 2025, with figures from IRS Publication 501 (2025) and the 2025 Instructions for Schedule A. It is exam preparation, not tax advice.

Star schema and relationships in Power BI

Fri, 09 Oct 2026 00:00:00 +0000

A star schema sorts every table in a semantic model into one of two kinds: dimension tables that describe things and are used to filter and group, and fact tables that record events and are used to summarize. The decision rule for exam questions: relate each dimension to the fact table with a one-to-many relationship, filtering in a single direction from the dimension (one side) to the fact (many side), and change that only when the scenario forces it. Microsoft’s guidance puts it as dimension tables enabling filtering and grouping, and fact tables enabling summarization.

Static vs dynamic routing: when to use which

Fri, 09 Oct 2026 00:00:00 +0000

The decision in one paragraph

A static route is typed in by an administrator and never changes unless someone edits it; a dynamic route is learned from a routing protocol such as OSPF, EIGRP, RIP or BGP and updates itself when the topology changes. Use static routes where the path is fixed and small: a stub site with one exit, a default route to an ISP, a host route to one server. Use dynamic routing where there are many routers, several possible paths or frequent change. The exam rule: if the scenario says the network must adapt automatically to a failure, the answer is dynamic; if it says low overhead, a single path or tight control, the answer is static.

Subnetting practice problems with answers

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free subnetting practice problems with explained answers, from finding a network address to VLSM planning and route summarisation. Subnetting is the skill of splitting an IPv4 network into smaller ones by moving the prefix boundary, and every answer here was recomputed with code. It is tested on the CCNA 200-301, where configuring and verifying IPv4 addressing and subnetting is an explicit objective, and in simpler form on CompTIA Network+ N10-009.

How to use these questions

Work each problem on paper before opening the answer. Write the block size first (256 minus the interesting mask octet), because most problems reduce to counting in multiples of it. Note the constraint word in the stem: “usable”, “smallest”, “exactly” and “broadcast” each change the answer. The method behind every step is in subnetting step by step.

Subnetting step by step, with worked examples

Fri, 09 Oct 2026 00:00:00 +0000

The one rule behind every subnetting question

Subnetting means splitting an address into a network part and a host part, and every question asks for the same few numbers: the network address, the broadcast address, the usable host range or the number of hosts. The fastest reliable method is the block size: find the octet where the mask stops being 255 or 0, subtract that mask value from 256, and the result is the step between subnets in that octet. Network addresses are the multiples of the block size, and the broadcast address is one below the next multiple.

Switch, router, firewall, IDS and IPS compared

Fri, 09 Oct 2026 00:00:00 +0000

Each device is defined by what it reads in the packet

A switch forwards frames inside one network using MAC addresses. A router forwards packets between networks using IP addresses. A firewall permits or blocks traffic by policy, based on addresses, ports, connection state or application. An IDS (intrusion detection system) watches traffic and raises alerts; an IPS (intrusion prevention system) sits in the path and can also drop the traffic. The decision rule for exam questions: find the layer the requirement talks about (MAC, IP, port or application, malicious pattern) and pick the device whose job is that layer, then check whether the requirement says “detect” or “block”.

Tax basis: how to find it for any asset

Fri, 09 Oct 2026 00:00:00 +0000

Basis is the amount of investment a taxpayer has in property for tax purposes, and it decides how much gain or loss is recognized on a sale. The decision rule: ask first how the taxpayer got the asset (bought, gifted, inherited, received for services, exchanged, converted from personal use), apply the starting-basis rule for that route, then adjust for improvements, depreciation and other events. This guide uses tax year 2025 and the IRS sources Publication 551 (12/2025), Publication 550 (2025), Publication 541 (12/2025) and the Instructions for Form 7203. It is exam preparation, not tax advice.

Unity Catalog explained: grants and storage

Fri, 09 Oct 2026 00:00:00 +0000

Unity Catalog is the governance layer built into Azure Databricks: it controls who can use which tables, files and AI assets, records lineage and audit logs, and tracks every asset under one three-level name, catalog.schema.object. The decision rule for exams: privileges are granted at a level and inherited downward, and reading a table needs three things (USE CATALOG, USE SCHEMA, SELECT), while files are governed through volumes, not through raw storage paths. Most Unity Catalog questions are a variation on one of those two sentences.

VLAN and trunking practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free VLAN and trunking practice questions with explained answers, written for CCNA candidates. A VLAN splits one switch into separate Layer 2 broadcast domains, and a trunk carries several VLANs between devices with 802.1Q tags; traffic between VLANs needs routing. The topic is tested in the Network Access domain of CCNA 200-301. All questions are original and use Cisco IOS behavior.

How to use these questions

Try each question before opening the answer, and commit to one letter. Underline the constraint word in the stem, such as “untagged”, “same VLAN”, “different VLANs” or “one physical link”, because it usually points to the right mechanism. Read every explanation, including for questions answered correctly, since each one says why the other options fail. The background is in the VLANs and trunking topic guide.

VLANs and trunking: 802.1Q, native VLAN, routing

Fri, 09 Oct 2026 00:00:00 +0000

What a VLAN does, and the rule that solves most questions

A VLAN (virtual LAN) splits one physical switch into separate Layer 2 broadcast domains, so devices in different VLANs cannot reach each other without a router, even when they are plugged into the same switch. A trunk is a link that carries several VLANs at once between switches, or between a switch and a router, by adding an 802.1Q tag that names the VLAN to each frame. The rule that answers most exam questions: an access port belongs to exactly one VLAN and sends frames untagged; a trunk port carries many VLANs, tagged; and traffic between two VLANs always needs a Layer 3 device, either a router or a Layer 3 switch.

VPN types and IPsec: AH, ESP and IKE

Fri, 09 Oct 2026 00:00:00 +0000

Pick the VPN type from who connects and what they reach

A VPN builds an encrypted tunnel across a network you do not trust, usually the internet, so that the two ends behave as if they were on one private network. The decision rule: if the requirement connects two networks permanently, it is a site-to-site VPN between gateways; if it connects one user or device to a network, it is a client-to-site (remote access) VPN. Then look at the two refinements the exam likes, clientless and split tunnel. IPsec is the protocol suite that most site-to-site VPNs, and many remote access VPNs, use underneath.

Vulnerability management and CVSS scoring

Fri, 09 Oct 2026 00:00:00 +0000

Vulnerability management is the repeating process of finding weaknesses in systems, ranking them, fixing or mitigating them, and confirming the fix. CVSS (Common Vulnerability Scoring System) is the standard way to express how severe a vulnerability is, on a scale of 0.0 to 10.0. The decision rule for the exam: CVSS measures severity of the flaw itself, not risk to your organisation, so the first fix goes to the vulnerability that is both severe and exploitable on a system that matters, not simply to the highest number.

Wireless networking: 802.11, bands, WPA3, APs

Fri, 09 Oct 2026 00:00:00 +0000

The four choices every wireless question hides

Almost every wireless question on Network+ and the CCNA reduces to four decisions: which standard, which band, which security mode and which management model. The rule of thumb is that range and speed trade against each other (lower frequency goes farther, higher frequency carries more), and that security is chosen by who needs to log in (one shared passphrase versus individual accounts). This guide takes the four decisions in turn, with a table for each.

YAML pipelines basics: Azure and GitHub

Fri, 09 Oct 2026 00:00:00 +0000

A YAML pipeline is a text file in your repository that tells a CI/CD service when to run, where to run, and which steps to execute. Azure Pipelines and GitHub Actions use different keywords but the same shape: a trigger starts a run, the run is split into jobs, and each job is an ordered list of steps on one machine. The exam rule: jobs are independent machines that run in parallel unless you order them, steps share one machine and run in sequence.

Zero Trust and IAM practice questions

Fri, 09 Oct 2026 00:00:00 +0000

Fifteen free Zero Trust and identity and access management (IAM) practice questions with explained answers. Zero Trust is a security model that grants no implicit trust based on network location and verifies every request using identity, device health and context. IAM covers how accounts are provisioned, authenticated, authorized and reviewed. Both are tested in Security+ SY0-801.

How to use these questions

Answer before reading the explanation, then check why each wrong option fails.
Note the constraint word in each stem: “least”, “first”, “most”, or the specific control type being asked for.
Separate authentication (who are you), authorization (what may you do) and accounting (what did you do).
The topic guide Zero Trust explained covers the principles and the policy decision and enforcement split.

Warm-up: principles and access models

Question 1. Which statement best describes the Zero Trust model?

Zero Trust explained for security exams

Fri, 09 Oct 2026 00:00:00 +0000

Zero Trust is a security model that drops the idea of a trusted internal network: every request is verified on its own evidence, whatever network it comes from. Microsoft states it as three principles, verify explicitly, use least privilege access and assume breach. The decision rule for the exam: if a scenario trusts a user or device because it is on the corporate network or behind the VPN, that is the old model; if the answer checks identity, device health and context on every request, grants only what is needed and limits damage if something is compromised, that is Zero Trust.

Udemy coupons for the practice tests

Fri, 02 Oct 2026 00:00:00 +0000

Below you find every practice test course we publish on Udemy. When coupons are running, they are listed in a separate box at the top of the course list, with the date they end; the coupon is already in the link, so Udemy shows the discounted or free price when you click. When no coupon is running, that box does not appear and the courses are available at their regular Udemy price.

Certification exam changes tracker

Sun, 20 Sep 2026 00:00:00 +0000

Certification bodies revise exam objectives, retire versions and replace whole exams, often with limited notice. This page tracks the changes affecting the exams covered on this site, so you find out before you book rather than after.

Last reviewed: 29 September 2026.

Recent and confirmed changes

Exam	Status	What this means for you
AI-103	Replaced AI-102	AI-102 was retired on 30 June 2026. Anything written for AI-102 is out of date — the weighting moved substantially towards agents, RAG and running AI in production.
AI-200	Replaced AZ-204	AZ-204 and the Azure Developer Associate certification were retired on 31 July 2026, including the renewal assessment. The storage domain was replaced by vector search in Cosmos DB, PostgreSQL and Redis, so AZ-204 material is out of date.
AI-300	Replaced DP-100	DP-100 and the Azure Data Scientist Associate certification were retired on 1 June 2026, including the renewal assessment. AI-300 adds MLOps with GitHub Actions and Bicep plus GenAIOps in Foundry, so DP-100 material covers only part of it.
AZ-802	Replaces AZ-800 and AZ-801	AZ-800 and AZ-801 retire on 30 September 2026; from then on AZ-802 alone earns the certification, renamed Windows Server Administrator Associate. Existing holders keep it. AZ-802 drops containers, clustering, migration, backup and Site Recovery, so AZ-800/801 material covers much more than you need.
AI-901	Restructured in 2026	Rebuilt around Microsoft Foundry. Older material covers much the same services under different framing, so check the publication date.
AB-900	New exam, outline update 14 October 2026	The fundamentals exam for Microsoft 365 administrators who manage Copilot and agents. MS-900 retired on 31 March 2026. The English outline changes on 14 October 2026 (a minor change to the core security features group), so check which version your material follows.
N10-009	Current version	Superseded N10-008. Material for the old version still circulates widely — check which version your study material and your voucher are for.
MLA-C02	Current version	The version being offered now. Do not use material aimed at the retired ML Specialty exam, which was far more theory-heavy.
CY0-001	New exam	Recently introduced. Third-party material is thin and some of it was adapted from generic AI security courses rather than written to the objectives.
AAIA, AAIR, AAISM	New certificates	Recently introduced by ISACA. Official material is effectively the only source written to the syllabus.
Agentforce Specialist	Current version	The version being offered now. Trailhead tracks it; older third-party courses may predate the current section weights.
SEE Part 1, Part 2, Part 3	New vendor, new fee	PSI replaced Prometric in 2026, the fee rose from 267 to 317 USD per part, and the exam now has three locked sections. The window running to 28 February 2027 tests tax law as amended through 31 December 2025. No testing in March and April.

No exam covered on this site is currently announced for retirement. Where a row says “current version”, it means that exam is the one being offered today and no successor has been announced — not that it will never change.

Certification exam glossary

Sun, 20 Sep 2026 00:00:00 +0000

Definitions of terms that recur across the exams covered on this site, written the way the exams actually use them rather than the way a textbook would.

AI and generative AI

Foundation model. A large model trained on broad data and adapted to many tasks, rather than built for one. Its generality is exactly what makes its behaviour hard to enumerate and test.

Embedding. Text represented as a vector of numbers, positioned so that similar meanings sit close together. This is what makes semantic similarity search possible. Not compression, not encryption.

Exam day: what to expect

Sun, 20 Sep 2026 00:00:00 +0000

Most people book their first proctored exam without knowing what the room check involves, what identification is accepted, or what happens if they fail. None of that is complicated, but discovering it fifteen minutes before your exam is a bad time. This page covers what is common across Microsoft, AWS, Databricks, CompTIA, ISACA and Salesforce.

Online or test centre

Every vendor covered here offers both. The trade-off is real.

Online proctored is convenient and strict. A person or software watches you through your webcam for the whole exam. Your room must be clear, your desk must be visible, and you cannot leave the frame — in most cases not even for a bathroom break.

How to read a skills outline

Sun, 20 Sep 2026 00:00:00 +0000

The skills outline is the document each exam is written from. It is free, it is published by the certification body, and it is the only authoritative statement of what can appear. Most candidates glance at it once and then study from a course written by someone else’s reading of it. That is where preparation goes wrong.

Why it beats every other source

Course material lags. A vendor reweights a domain, and the courses take months to catch up — some never do. Study guides summarise, and summarising means deciding what to leave out on your behalf, without knowing which domain you are weak in.

What certification exams cost in 2026

Sun, 20 Sep 2026 00:00:00 +0000

Exam fees for the certifications covered here run from 75 to 599 US dollars, and several vendors do not publish a price at all. This page lists what each one actually costs, and is explicit about where a figure is unavailable rather than inventing one.

Published fees

Exam	Fee	Notes
MLA-C02	75 USD	Beta pricing; expect it to rise once the beta ends
AIF-C01	100 USD	Foundational level
AI-103	165 USD	US price; varies by country
Agentforce Specialist	200 USD
Data Analyst	200 USD
Data Engineer Associate	200 USD
Data Engineer Professional	200 USD	Same fee as the associate
GenAI Engineer	200 USD
ML Associate	200 USD
ML Professional	200 USD
Spark Developer	200 USD
AAIA	459 USD member / 599 USD non-member	Plus a 50 USD application fee after passing
AAIR	459 USD member / 599 USD non-member	Plus application fee
AAISM	459 USD member / 599 USD non-member	Plus application fee
SEE Part 1, Part 2, Part 3	317 USD per part	951 USD for all three, plus a 140 USD enrollment fee (Form 23) after passing

Where no price is published

This is not an oversight on our part. These vendors genuinely do not put a figure on the certification page.

About ExamSprintLab

Mon, 01 Jan 0001 00:00:00 +0000

ExamSprintLab publishes study guides and practice exams for IT and tax certifications. The practice exams are written by an IT infrastructure specialist with more than 15 years of experience in cloud, DevOps and infrastructure.

How the guides are made

Every guide starts from the certification body’s own exam page and skills outline, the free document each exam is written from. Facts such as duration, price, passing score and domain weights are only published once they are confirmed in an official source. Where a number is not published, the guide says so instead of guessing. The guides are written in our own words and are not copied from vendor documentation or from other sites.

Privacy

Mon, 01 Jan 0001 00:00:00 +0000

This site does not use cookies. The only personal data we ask for is an email address, and only if you choose to subscribe to exam updates.

Analytics. We use Cloudflare Web Analytics to count page views. It does not set cookies and does not track you across sites.

Hosting. The site is hosted on Cloudflare Pages. Cloudflare processes technical data such as IP addresses to deliver and protect the site.