AIF-C01 study scenarios: three profiles
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.
Example profiles, not real people. The profiles below are illustrative planning examples built from the official exam outline. They are not reviews or reports of anyone’s result.
The site’s study plan assumes some AWS familiarity and no AI background, at 5 to 7 hours a week for four weeks. The profiles below move hours around that baseline. All hour figures are planning assumptions, not predictions.
The developer who already calls managed AI APIs
Starting point. Comfortable with IAM, S3 and Lambda, and has sent prompts to a managed model. Domain 3 feels familiar in practice (prompt techniques, inference parameters, retrieval), but the exam asks for the concept and the service name, not the code.
Where the time goes. Domain 1, fundamentals of AI and ML (20%), is the surprise: terminology such as precision versus recall, types of inferencing and when AI is not appropriate. Domain 4, responsible AI (14%), and domain 5, security, compliance and governance (14%), are rarely part of an API-calling day. Domain 2 (24%) also needs attention for the version 1.1 additions: token-based pricing, context engineering and agentic AI concepts including the Model Context Protocol.
The trap. Over-studying implementation. Anything about coding models, tuning or building pipelines belongs to MLA-C02 and is not examined here.
Order. Domain 1, then domain 4 and 5 while the vocabulary is fresh, then domains 2 and 3 as a quicker pass.
Time budget. A reasonable planning assumption is 14 to 18 hours over two to three weeks: about 2 to 3 on domain 1, 3 to 4 on domain 2, 4 to 5 on domain 3, 3 to 4 on domains 4 and 5 together, plus two hours of practice.
The business-side candidate with no AWS hands-on
Starting point. Scopes, buys or sells AI work. Understands what a chatbot does, but has not opened Amazon Bedrock or SageMaker AI, and the recommended knowledge (EC2, S3, Lambda, IAM, the shared responsibility model, pricing models) is partly new.
Where the time goes. Domain 3, applications of foundation models (28%), is the largest and the most technical: selection criteria, retrieval-augmented generation, vector databases, fine-tuning versus in-context learning, and evaluation metrics. Domain 2 (24%) supplies the language for it. Because the exam is compensatory, the 14% domains cannot carry a weak 28% domain on their own.
The trap. Treating domain 3 as too technical to bother with and leaning on the two governance domains, which together are only 28%.
Order. Follow the study plan’s weighting: domain 1, domain 2, domain 3, then domains 4 and 5, with an AWS account open from week 1 so service names stick.
Time budget. A reasonable planning assumption is 26 to 34 hours over five weeks: the plan’s 22 to 30 hours plus a few for the AWS basics. Practise ordering and matching questions specifically, because they are all-or-nothing.
The analyst or ML-curious candidate heading toward MLA-C02
Starting point. Knows supervised and unsupervised learning, features and labels, and has trained a model. Domain 1 is mostly known. Foundation models and agent-era vocabulary are not.
Where the time goes. Domains 2 and 3 (52% combined), especially the cost trade-offs between pre-training, fine-tuning, in-context learning, RAG and distillation. Domain 4 includes the effects of bias and variance, which is familiar, but also legal risks of generative AI and transparency tooling such as model cards.
The trap. Answering at an engineer’s depth. A question that asks which approach is cheapest for a company-specific knowledge base wants the high-level choice, not a training recipe.
Order. Skim domain 1 for AWS service mapping, then domain 3, domain 2, and finish with domains 4 and 5.
Time budget. A reasonable planning assumption is 15 to 19 hours: 1 to 2 on domain 1, 4 to 5 on domain 2, 5 to 6 on domain 3, 3 to 4 on domains 4 and 5, plus two hours of practice.
The three AIF-C01 profiles compared
| Profile | Front-load | Main risk |
|---|---|---|
| Developer calling managed AI APIs | Domain 1 (20%), domains 4 and 5 (14% each) | Studying implementation depth that is out of scope |
| Business-side, no AWS hands-on | Domain 3 (28%) and domain 2 (24%) | Skimming the biggest domain and AWS service names |
| ML-curious analyst | Domains 2 and 3 (52%) | Answering at engineer depth and missing 1.1 additions |
How to check which AIF-C01 profile fits before booking
Take the free sample questions cold and note which domain each wrong answer belongs to. If the misses cluster in domains 1, 4 and 5, you are closest to the developer profile; if they cluster in domain 3, you are closest to the business-side one. Then read the exam format for the compensatory scoring rule and the 15 unscored questions, and follow the study plan with your adjusted hours. Close with the free 20-question practice test, then a full timed practice exam at the real length: 65 questions in 90 minutes.
If you start from a newcomer position in IT, the career page notes AWS’s own suggestion to take a cloud essentials course first. If you are weighing the next step up, the MLA-C02 exam hub shows what AWS expects there.