Skip to content

Assembling and Deploying Apps Questions

Practice questions for Assembling and Deploying Apps topic in Databricks Certified Generative AI Engineer Associate. 44 questions covering this domain.

44 questions12 easy21 medium11 hard
Q1
easy

What is the Databricks AI Playground used for?

Q2
medium

An engineer uses code-based logging to deploy an AI agent with `mlflow.langchain.log_model(lc_model="/path/to/agent.py", ...)`. When the deployed serv...

Q3
medium

A Databricks admin wants to audit LLM token consumption by user for cost attribution. Unity AI Gateway usage tracking is enabled. Which two system tab...

Q4
medium

A team deploying a shared LLM endpoint wants to prevent any single user from consuming all endpoint capacity. Which Unity AI Gateway feature enables p...

Q5
easy

Before deploying an AI agent to a Mosaic AI Model Serving endpoint, which step is required to enable Unity Catalog governance and authorization for th...

Q6
medium

A Model Serving endpoint has two external models listed: Model A (60% traffic, listed first) and Model B (40% traffic, listed second). Unity AI Gatewa...

Q7
medium

A team enables PII detection guardrails on a Model Serving endpoint with the "Mask" option. A user sends a request containing a credit card number. Wh...

Q8
hard

An ML engineer has logged and registered an AI agent to Unity Catalog. Before deploying to a production serving endpoint, they want to validate the ag...

Q9
easy

An engineer enables payload logging (inference tables) via Unity AI Gateway on a model serving endpoint. How soon after a request is made can the team...

Q10
medium

A team enables traffic splitting with Model A at 20% and Model B at 80% on a single endpoint with fallbacks enabled. Model A is listed first in the en...

Q11
hard

A team is deciding between a service principal token and a personal access token (PAT) for authenticating a production AI agent serving endpoint to Ve...

Q12
easy

What does the Unity AI Gateway's "AI Guardrails" feature provide in the context of a Model Serving endpoint?

Q13
hard

An enterprise team has deployed an AI assistant serving endpoint used by 500 employees. They need to attribute LLM token costs to individual business ...

Q14
medium

A team configures a Unity AI Gateway rate limit on a Model Serving endpoint. They want individual users to be limited to 60 queries per minute each, b...

Q15
hard

A team manages a multi-tenant AI assistant where requests from different enterprise clients must be audited separately for compliance. Each client's r...

Q16
medium

A model serving endpoint has three external models listed in order: ModelX (50% traffic), ModelY (30% traffic), ModelZ (20% traffic). Unity AI Gateway...

Q17
easy

After registering an AI agent in Unity Catalog, how is a Model Serving endpoint created to deploy the agent using the Databricks `databricks.agents` P...

Q18
medium

A team deploys an AI agent and wants to ensure that topic guardrails prevent the endpoint from answering questions about competitor products. Which Un...

Q19
hard

An engineering team is migrating a production AI agent from using a personal access token (PAT) to a service principal for authenticating to Vector Se...

Q20
easy

When an engineer enables inference tables (payload logging) on a Model Serving endpoint, where is the logged request and response data stored?

Sign in to see all 44 questions

Create a free account to browse all questions — completely free during our launch phase.