Skip to content

Databricks Certified Generative AI Engineer Associate Flashcards

20 cards from this guide's fact sheet and notes. Click a question to reveal the answer.

To study with spaced repetition, download flashcards.csv and import it into Anki or Quizlet (comma separated, first row is the header). Back to the study guide.

Exam logistics

How long is the Databricks Certified Generative AI Engineer Associate exam?

90 minutes

How many questions are on the Databricks Certified Generative AI Engineer Associate exam?

45 scored multiple-choice or multiple-selection items

What is the passing score for Databricks Certified Generative AI Engineer Associate?

70% (approximately 32 correct)

What does the Databricks Certified Generative AI Engineer Associate exam cost?

$200 USD

How is the Databricks Certified Generative AI Engineer Associate exam delivered?

Online proctored

How long is Databricks Certified Generative AI Engineer Associate valid for?

2 years

Design applications and agent bricks

Knowledge Assistant (design applications and agent bricks)

takes a set of documents and produces a question-answering chat agent that cites its sources; it does not fill a table or take actions. The Agent Bricks FAQ now lists it as legacy, pointing new low-code agents to Genie Agents.

Information Extraction (design applications and agent bricks)

takes unstructured documents or text and returns structured fields that follow a schema you define, as JSON.

Multiagent Supervisor (design applications and agent bricks)

coordinates several existing agents and tools and routes each part of a request to the right one; the docs call it Supervisor Agent and the FAQ lists it as legacy.

Genie space (design applications and agent bricks)

a curated natural-language interface over a set of Unity Catalog tables: you ask in plain English, Genie writes and runs the SQL, and returns the answer.

Assembling deploying and mcp

Vector Search endpoint (assembling deploying and mcp)

the compute that serves vector queries (the docs now say AI Search endpoint, formerly Vector Search endpoint); an index lives on an endpoint.

Delta Sync index (assembling deploying and mcp)

a Vector Search index kept in step with a source Delta table, with embeddings computed by Databricks or supplied by you (self-managed).

Direct Vector Access index (assembling deploying and mcp)

a Vector Search index that is not synced automatically: you update it yourself through the REST API.

Prompt Registry (assembling deploying and mcp)

the MLflow tool that stores a prompt as a named Unity Catalog object and creates an immutable, numbered version each time it changes.

Managed MCP server (assembling deploying and mcp)

an MCP server provided and hosted by Databricks that exposes platform capabilities, such as Databricks SQL and Genie One, as tools.

Custom MCP server (assembling deploying and mcp)

an MCP server you build, typically from Databricks Apps templates, to expose your own logic or wrap an internal REST API.

Evaluation monitoring and governance

Inference tables (evaluation monitoring and governance)

log each request and response sent to a serving endpoint into a Unity Catalog Delta table.

Usage tracking (evaluation monitoring and governance)

records requests, tokens and latency per endpoint in system tables, so you can see who is using what and what it costs.

Production monitoring (evaluation monitoring and governance)

runs your scorers automatically on a sample of production traffic, so quality is checked continuously after launch.

Ground truth (evaluation monitoring and governance)

an expected answer or expected facts supplied with each evaluation record; Correctness, RetrievalSufficiency and ToolCallCorrectness need it.