Databricks GenAI Engineer exam format
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.
Multiple choice only. No drag-and-drop, no ordering, no matching. That makes the format simpler than most AWS or Microsoft exams.
The passing score is not published
Databricks does not publish a passing score for this exam. Treat any specific figure you find elsewhere with caution unless it cites Databricks directly. Aim for consistent 80%+ on realistic practice material rather than calculating a margin.
Section weighting
| Section | Weight | Approx. scored questions |
|---|---|---|
| Design Applications | 14% | ~6 |
| Data Preparation | 14% | ~6 |
| Application Development | 30% | ~14 |
| Assembling and Deploying Apps | 22% | ~10 |
| Governance | 8% | ~4 |
| Evaluation and Monitoring | 12% | ~5 |
Counts are derived from Databricks’ percentages against 45 scored questions.
Two observations worth acting on:
Building and shipping is 52%. Application development plus assembling and deploying. If your time is limited, this is where it goes.
Governance is only 8% — roughly four questions. It is the smallest section on the exam. Learn it, because four questions matter at the margin, but do not spend a week there.
Two-year validity
Shorter than AWS’s three years, longer than the annual cycle on Microsoft associate certifications and ISACA credentials. Recertification is required to keep the credential current, so factor in a recurring commitment.
What the exam is really about
Databricks describes the certification as testing the full lifecycle of building generative AI applications on the platform. In practice that means RAG applications and LLM chains are the centre of gravity: how you chunk and index content, how you build a chain, how you deploy it, and how you know whether it works.
If your preparation is mostly conceptual reading about large language models, you are studying for a different exam. This one rewards having built something on Databricks.
Booking
Book through Databricks. No prerequisites, so nothing gates registration — but the 6+ months of hands-on experience Databricks recommends is a realistic expectation rather than a formality.