Generative AI Engineer Associate: practice questions
Original practice questions for building GenAI systems on Databricks: vector search, retrieval, agents, serving LLMs, and evaluation. Citation-grounded in the current docs, with the source linked next to every answer.

GenAI Engineer set
12 questions, drawn fresh from the bank. Every code question was executed on Spark before it shipped, and each answer links to the documentation it rests on.
No account needed to practise. Sign in to keep your best score per set, see progress on your account page, and turn wrong answers into a reading list.
A sample question, with the answer
A Delta Sync vector index's source table receives new rows. What keeps the index current?
- aNothing; you rebuild the index on a schedule you own
- bA trigger you must write on the source table
- cThe index syncs with the source Delta table automaticallycorrect
- dOnly OPTIMIZE on the source table refreshes it
Why: A Delta Sync index tracks its source Delta table and updates as the table changes, continuously or on triggered syncs depending on configuration. That is the point of choosing it over a Direct Access index, where you manage updates yourself.
Exam facts
Facts reflect the official materials at our last review; exams change, so verify against the official page before booking.
Prep with the AI assistant you already use
Databricks’ recommended prep flow now includes an official AI Prep Guide of copy-paste prompts. Here is ours, tuned to this exam. Paste it into any assistant, or ask the Brickster Assistant, which answers with citations from this archive.
I am preparing for the Databricks Certified Generative AI Engineer Associate exam. Quiz me with 10 original multiple-choice questions in the style of that exam, one at a time, across these areas: agents, vector-search, mlflow, genie. Never reuse real exam content. Do not show the answer until I respond. After each answer, explain why, and ground the explanation in docs.databricks.com. Keep score. At the end, list my weakest areas and what to read for each.
Go deeper on the weak spots
Each topic below has a live page with an AI synthesis of recent activity and the items behind it. Wrong answers in the set above link to these too.
Frequently asked
How many questions is the Generative AI Engineer Associate exam?
45 scored questions in 90 minutes across six domains, per the official exam guide dated March 2026. It costs $200, is proctored, and the credential is valid for two years. Databricks does not publish the passing score.
Are these real exam questions?
No. Every question is original, written from public documentation with the supporting page cited next to the answer. Real exam content is covered by the exam NDA, and Databricks warns that dumps can void a certification.
What should I study beyond practice questions?
The official exam guide lists the six domains, and the official AI Prep Guide provides copy-paste prompts for any assistant. Our topic hubs for vector search, agents, MLflow, and Genie carry syntheses of what is changing, which matters for an exam refreshed as recently as March 2026.
brickster.ai is an independent community project, not affiliated with, endorsed by, or sponsored by Databricks, Inc. Nothing here is drawn from real exam content, and no practice result predicts a certification outcome.
