Databricks-GenAI-Assoc Practice Test 12 — 15 Questions

Question 1

Which Databricks feature is specifically designed to allow developers to programmatically manage and version their entire data and AI infrastructure as code?

Accepted Answer

Databricks Asset Bundles. Databricks Asset Bundles (DABs) enable developers to package and deploy their projects—including jobs, pipelines, and models—using a declarative, version-controlled approach. This is fundamental for modern MLOps and DataOps workflows, as it replaces manual UI configuration with repeatable, automated deployments. By treating infrastructure as code, teams improve reliability, auditability, and speed when promoting applications from development to production environments.

Answer

Unity Catalog Volumes

Answer

SQL Warehouses

Answer

Delta Sharing

Question 2

An AI engineer is developing a real-time customer service chatbot application using Databricks Model Serving and needs to securely store API keys and database credentials without hardcoding them into the application code or notebook. Which approach should the engineer use?

Accepted Answer

Utilize Databricks Secrets to create a secret scope and retrieve values securely using dbutils.secrets.get during application execution.. Databricks Secrets provide a secure mechanism for managing credentials and sensitive configuration parameters. By leveraging secret scopes and the dbutils.secrets utility within notebooks or environment variables mapped to serving endpoints, developers ensure that sensitive tokens never leak into version control systems, adhering to strict enterprise security and compliance standards for production application deployments.

Answer

Store the credentials in a plain-text JSON file located in the workspace root directory and read it programmatically at runtime.

Answer

Hardcode the credentials directly into the scoring script deployed to the Databricks Model Serving endpoint.

Answer

Pass the credentials as plain-text arguments in the command-line interface when triggering the job cluster.

Question 3

An organization is deploying an LLM application using Databricks Asset Bundles (DABs). Which TWO of the following are primary benefits of using DABs for the deployment process?

Accepted Answer

Enables infrastructure-as-code (IaC) for Databricks resources.. Databricks Asset Bundles streamline the development and deployment lifecycle by treating infrastructure and code as a single unit. They facilitate consistent environments across development, staging, and production by using YAML-based configuration files. This ensures that the deployment process is reproducible, automated via CI/CD pipelines, and manageable through version control, reducing human error and configuration drift in complex ML-driven applications.

Answer

Automatic generation of unit tests for Python code.

Answer

Provides a drag-and-drop interface for building pipelines.

Answer

Automatically migrates all legacy data to Unity Catalog.

Question 4

When deploying a Python-based application that interacts with Unity Catalog, which step is essential to ensure the code can authenticate securely to external services without hardcoding tokens?

Accepted Answer

Using Databricks Secret Scopes to retrieve credentials.. Using secret scopes within Databricks is the recommended best practice for handling authentication credentials. Secrets allow code to retrieve sensitive information like API keys or database passwords at runtime, keeping these credentials out of the source code. This is a critical security requirement in any production application, as it prevents accidental exposure of sensitive keys in version control systems and allows for centralized management of authentication lifecycle.

Answer

Hardcoding tokens as environment variables in the notebook.

Answer

Storing tokens in a plain text file inside the bundle.

Answer

Creating a public access policy for the external service.

Question 5

A team deploys a customer-support assistant on a Mosaic AI Model Serving endpoint and enables inference table logging to Unity Catalog. Compliance requires that every production response be traceable back to the exact request, the retrieved context, and the model version that produced it, and that reviewers can query this history with SQL months later. Which capability satisfies this requirement?

Accepted Answer

The endpoint's request and response payloads written to a Unity Catalog inference table, queried with SQL and joined to model version metadata.. Compliance-grade traceability demands durable, queryable records of each request, the context supplied to the model, the response, and the serving model version. Unity Catalog inference tables store exactly that as governed Delta tables, so SQL queries and joins to model version metadata answer audit questions months after the traffic occurred. Client logs, offline experiment runs, and aggregate dashboards all lack either the payload detail or the retention and governance needed.

Answer

Real-time dashboards built on the endpoint's built-in metrics such as request count and error rate.

Answer

MLflow experiment runs recorded during offline evaluation of the assistant before release.

Answer

Client-side application logs streamed to a log analytics workspace and retained for thirty days.

Question 6

An engineer built an agent using Mosaic AI Agent Framework and wants the agent to call a Unity Catalog function that returns customer order history. The function must be invoked by the LLM at runtime without exposing raw SQL to the model. Which approach should the engineer use?

Accepted Answer

Register the Unity Catalog function as a tool in the agent and let the LLM select it by name and arguments.. Registering the Unity Catalog function as a tool gives the agent a typed, governed interface the LLM can invoke by name with structured arguments. The function's metadata supplies the schema, so no SQL is exposed to or generated by the model. Prompt-embedded SQL, scheduled jobs, and text-to-SQL all fail the requirement of runtime, governed invocation without model-authored queries.

Answer

Expose the function through a SQL warehouse and have the agent send natural-language queries to the warehouse endpoint.

Answer

Embed the SQL body of the function in the system prompt so the LLM can generate the query when needed.

Answer

Create a Databricks job that runs the function on a schedule and writes results to a Delta table the agent reads.

Question 7

A GenAI engineer is preparing a large text corpus stored in a Unity Catalog volume for fine-tuning a chat model. The raw files are JSON Lines, each containing a 'conversation' array with alternating 'user' and 'assistant' turns, but many records have malformed turns or missing roles. The engineer needs to convert this into a Delta table with a schema of (conversation_id STRING, messages ARRAY>) while filtering out records where any turn has a null role or empty content. Which approach uses the appropriate Databricks-native capability for this transformation?

Accepted Answer

Use the from_json function with a defined StructType schema to parse each JSON line into the conversation structure, then apply a filter using a higher-order function such as filter on the messages array to retain only records where all turns have non-null role and non-empty content.. The correct approach uses from_json with an explicit schema to parse the nested JSON into the required array of structs, then applies a higher-order function to filter out records with malformed turns. This leverages Spark SQL's native, optimized JSON parsing and array operations, which scale well and enforce the target schema. It avoids the performance pitfalls of UDFs, the fragility of regex, and the indirectness of ingestion frameworks not designed for immediate nested transformation.

Answer

Load the files as a text RDD, apply a regular expression to extract role and content pairs, and then convert the RDD to a DataFrame with the desired schema using createDataFrame.

Answer

Use Databricks Auto Loader with schema evolution to ingest the files into a Bronze table, then use Delta Live Tables expectations to drop rows where the role column is null or the content column is empty.

Answer

Use the spark.read.json method with inferSchema enabled to automatically detect the schema, then use a UDF written in Python to iterate over each turn and remove records with null roles or empty content.

Question 8

A Generative AI engineer is deploying a new version of a RAG chain to a Mosaic AI Model Serving endpoint. Before promoting it to production, they want to run an evaluation that checks whether the generated answers are faithful to the retrieved documents. Which Mosaic AI Agent Evaluation metric should they examine?

Accepted Answer

groundedness. To verify that generated answers are faithful to the retrieved documents, the evaluation must compare the answer against the context. groundedness is the Agent Evaluation metric that performs this check, flagging responses that include unsupported claims. It is the appropriate pre-promotion quality gate for a RAG chain.

Answer

chunk_relevance

Answer

latency

Answer

request_count

Question 9

A GenAI engineer is designing a multi-turn chat application on Databricks. Users report that the assistant forgets details from earlier in long conversations and sometimes answers using only the most recent message. The team wants the model to reliably use facts stated several turns earlier without exceeding the model's context window. Which design should the engineer implement?

Accepted Answer

Implement a conversation memory strategy that summarizes or selectively retains earlier turns and injects the compressed history into each prompt within the context window.. Reliable multi-turn memory requires bringing earlier relevant turns into the prompt, either by summarizing them or by selectively retaining key facts. This keeps the payload within the context window while preserving details the user expects the assistant to remember. Temperature changes, sending only the latest message, or storing transcripts without retrieval all fail to put earlier facts in front of the model.

Answer

Store the full conversation transcript in a Delta table and instruct the model to query the table during inference.

Answer

Increase the model's temperature so it explores more of the conversation history when generating a response.

Answer

Send only the latest user message to the model and rely on the model's pretrained knowledge to recall earlier facts.

Question 10

A GenAI engineer is building a retrieval-augmented generation (RAG) application using Databricks Vector Search. During testing, they observe that for some queries, the retrieval step returns document chunks that are semantically similar but not actually relevant to the user's question, leading to poor answer quality. They want to improve retrieval precision without changing the embedding model. Which of the following approaches is most appropriate?

Accepted Answer

Enable hybrid search on the Vector Search index to combine keyword-based and vector similarity scores.. Hybrid search enhances retrieval precision by blending vector similarity with keyword matching, ensuring that chunks containing exact query terms are ranked higher. This directly addresses the problem of semantically similar but irrelevant results without altering the embedding model. The other options either do not tackle the precision issue or violate the constraint of not changing the embedding model.

Answer

Reduce the number of retrieved chunks (top_k) to only the single highest-scoring chunk.

Answer

Switch the embedding model to a larger, more powerful model to improve semantic representation.

Answer

Increase the chunk size of the documents to include more context per retrieved chunk.

Question 11

An engineer maintains a GenAI application that uses a Databricks Asset Bundle to deploy a Mosaic AI Agent serving endpoint. A new model version has been logged and validated, and the team wants to roll it out to production with the ability to revert quickly if quality regressions appear. Which deployment approach best satisfies this?

Accepted Answer

Create a new serving endpoint version by updating the endpoint's served model entity to the new model version, then verify quality and roll back by pointing the entity back to the previous version.. Updating the served model entity on an existing endpoint moves production traffic to the validated model version while keeping the endpoint URL and configuration stable, so clients are unaffected. If quality regresses, reverting the served entity to the previous version restores prior behavior quickly. Recreating endpoints, notebook-only testing, and cross-workspace routing each introduce downtime, delay, or unnecessary complexity instead of a clean, reversible cutover.

Answer

Keep the existing endpoint unchanged and run the new model version only in notebooks until the team is confident, then deploy it later.

Answer

Deploy the new model version to a second workspace and route production traffic there through a global load balancer.

Answer

Delete the production endpoint and create a brand-new endpoint with the new model version, then update all clients to the new URL.

Question 12

A GenAI engineer has a Mosaic AI Agent application packaged as a Databricks Asset Bundle with a serving endpoint defined in the bundle's resources. A teammate recently updated the agent's prompt template in the source files, and the engineer now needs to push that change to the existing production endpoint without recreating it. Which Databricks CLI command should the engineer run from the bundle root?

Accepted Answer

databricks bundle deploy --target prod. Deploying a modified Mosaic AI Agent that is defined as a bundle resource requires applying the bundle state to the target workspace, which the Databricks CLI does with bundle deploy. Selecting the production target ensures the correct workspace, variables, and permissions are used, and the existing serving endpoint is updated in place instead of being recreated.

Answer

databricks bundle validate --target prod

Answer

databricks serving-endpoints update --name agent-endpoint

Answer

databricks bundle run agent_app --target prod

Question 13

A data engineer is preparing a dataset of product descriptions for embedding generation. The source table in Unity Catalog contains a column 'description' with mixed languages, and the team wants to filter to English-only text before vectorization. They need a scalable, built-in Databricks function that can detect the language of each description without external API calls. Which function should they use?

Accepted Answer

ai_detect_language. The built-in ai_detect_language function is purpose-built for language identification at scale within Databricks. It returns a language code that can be used to filter the dataset to English-only records before embedding generation. Other AI functions like ai_analyze_sentiment or ai_classify serve different purposes, and regex cannot reliably determine language.

Answer

regexp_extract

Answer

ai_classify

Answer

ai_analyze_sentiment

Question 14

A GenAI team's Mosaic AI Agent application is deployed via a Databricks Asset Bundle and served through a Model Serving endpoint. They need the endpoint to call a Unity Catalog function as a tool during inference, and the function reads from a table the endpoint's service principal cannot currently access. What should the engineer do?

Accepted Answer

Grant the endpoint's service principal the required privileges on the table and on the Unity Catalog function, then redeploy the bundle.. Unity Catalog functions invoked as tools by a Model Serving endpoint execute under the endpoint's identity, which is a service principal. For the function to read its underlying table, that principal needs EXECUTE on the function plus the appropriate privileges on the table and its parent catalog and schema. Granting those privileges and redeploying the bundle resolves the access failure while keeping credentials out of the code.

Answer

Embed the service principal's personal access token in the agent code so the function can authenticate at runtime.

Answer

Add the function's source table to the endpoint's `auto_capture_config` so the endpoint can read it during inference.

Answer

Change the endpoint's `workload_type` to `GPU_LARGE` so the endpoint runs with elevated permissions.

Question 15

You are building a data preparation pipeline for a generative AI application. The raw data is stored in a Unity Catalog volume as JSON files with nested fields. You need to flatten the nested structure and extract specific fields into a Delta table for downstream embedding. The JSON schema may evolve, with new fields added occasionally. Which approach provides the most robust and maintainable solution?

Accepted Answer

Use the variant data type to store the JSON and then use variant_get to extract fields.. The variant data type stores JSON without a rigid schema, and variant_get extracts fields dynamically, accommodating schema evolution. This avoids brittle predefined schemas and manual updates. Other options either require static schemas, rely on sampling, or only handle arrays, making them less robust for evolving nested JSON.

Answer

Use the explode function on the nested arrays and then select the required fields.

Answer

Use the from_json function with a predefined schema and select the required fields.

Answer

Use the schema_of_json function to infer the schema at runtime and then apply from_json.