Databricks-GenAI-Assoc Assembling and Deploying Apps — Timed 20 Questions

Question 1

A machine learning engineer needs to deploy a Generative AI application using Databricks Model Serving. The application requires access to a private vector database within the same VPC. Which configuration step is mandatory for this deployment?

20 minute time limit — choose an answer to begin.

Accepted Answer

Configure a custom serving endpoint with a private network connection policy.. To enable Model Serving to communicate with private network resources like a VPC-internal database, the engineer must configure Private Link or a dedicated network connectivity setup. This ensures that traffic remains within the secure Databricks environment. Properly configuring these network settings is crucial for maintaining security compliance while allowing the model to perform retrieval-augmented generation (RAG) by querying internal data sources securely without exposing endpoints to the public internet.

Answer

Enable public endpoint access on the Model Serving cluster.

Answer

Deploy the model as a library within a standard Databricks notebook.

Answer

Increase the maximum instance count to ensure network path availability.

Question 2

Refer to the exhibit. An engineer wants to optimize cost while ensuring high availability for the 'genai-chat-bot' model. Given the current configuration, what is the most appropriate action?

Accepted Answer

Set min_instances to 0 to enable scale-to-zero capabilities.. The current configuration shows a static minimum instance count of one and a maximum of five with a GPU-based instance type. To optimize cost while maintaining availability, the engineer should consider scaling to zero when no traffic is detected. This strategy reduces costs significantly during idle periods while maintaining the ability to serve requests on demand, which is a best practice for managing expensive GPU-backed inference resources in Databricks.

Answer

Increase max_instances to 10 to improve throughput.

Answer

Change instance_type to CPU_MEDIUM to lower cost.

Answer

Delete the tags section to save configuration memory.

Question 3

Refer to the exhibit. An engineer observes that the training job takes too long to execute. Which modification to the DAB YAML configuration would be the most effective way to improve performance for a data-intensive training job?

Accepted Answer

Increase the 'num_workers' value in the cluster configuration.. Increasing the number of workers in the cluster configuration scales out the Spark processing power. For data-intensive training jobs, a larger cluster allows for more parallel processing of partitions, reducing overall runtime. This approach is the standard method for horizontal scaling within Databricks Jobs, ensuring the infrastructure can handle increased computational demand while maintaining the existing code logic.

Answer

Change the spark_version to an older stable release.

Answer

Rename the job to 'optimized_training_job'.

Answer

Remove the 'new_cluster' section and use the 'existing_cluster' ID.

Question 4

When managing multiple environments using Databricks Asset Bundles, how should an engineer manage environment-specific variables like database table names?

Accepted Answer

Define environment-specific variables in the 'targets' section of the YAML.. DABs support environment-specific overrides in the bundle configuration file. By defining a base configuration and then providing specific values for different environments (e.g., dev, staging, prod) in the workspace settings, the engineer can ensure the code uses the correct resources without manual modifications. This keeps the application code generic and portable across the entire development lifecycle.

Answer

Use hardcoded 'if-else' statements within the notebook code.

Answer

Create separate branches in Git for each environment.

Answer

Store environment variables in a Databricks SQL table.

Question 5

An engineer needs to ensure that a model serving endpoint deployment is secure and scalable. Which TWO of the following configurations should be prioritized in the DABs deployment file?

Accepted Answer

Enabling 'restricted' access via workspace-level network policies.. Securing and scaling model endpoints involves limiting access and managing resources. Enabling VPC-level network security ensures that the endpoint is not exposed to the public internet. Simultaneously, configuring autoscaling allows the endpoint to adjust capacity based on actual request volume, which prevents resource starvation during spikes. Together, these settings form the backbone of a production-ready, enterprise-grade deployment that is both secure and resilient.

Answer

Setting a fixed cluster size with 10 instances.

Answer

Disabling logging to increase inference throughput.

Answer

Using the 'default' public endpoint URL for testing.

Question 6

When using Databricks Asset Bundles to deploy a model, which command should an engineer use to view the status of the current deployment in the target environment?

Accepted Answer

bundle status. The 'bundle validate' command is for syntax checking, but 'bundle deploy' initiates the process. To check the status of existing resources within the bundle context, the Databricks CLI provides status introspection commands. Knowing how to monitor the state of deployed assets is essential for verifying that the CI/CD deployment was successful and that all resources are active and configured as expected in the production workspace.

Answer

bundle validate

Answer

bundle deploy

Answer

bundle list

Question 7

An engineer has a Databricks Job that runs hourly. They want to include this job in their Databricks Asset Bundle. What is the correct way to define this job in the bundle YAML?

Accepted Answer

Use the 'resources' -> 'jobs' section with the job configuration.. Defining a job within a bundle requires using the 'resources' section, specifically the 'jobs' key. Each job needs a unique key and a configuration block that specifies the task, compute, and schedule. This declarative approach allows the entire job definition, including dependencies and schedules, to be managed in version control. It ensures that whenever the bundle is deployed, the job configuration is synchronized with the Databricks environment.

Answer

Define the job in a separate Python script and call it from the YAML.

Answer

Create a notebook with the job schedule and import it.

Answer

Use the 'sync' section to automatically pull existing jobs from the workspace.

Question 8

An engineer has implemented a custom model in MLflow. Which component is required to serve this model using Databricks Model Serving?

Accepted Answer

The model must be registered in the Unity Catalog.. To serve a model, it must be registered in the Unity Catalog. Unity Catalog serves as the centralized model registry, providing the necessary metadata and lineage for serving endpoints to locate and deploy the correct model version. This governance layer is essential for production serving, as it ensures that the serving infrastructure has secure, authorized, and version-controlled access to the model artifacts.

Answer

A Docker container image of the entire Databricks workspace.

Answer

A local copy of the model stored as a pickle file in the Git repo.

Answer

A separate SQL Warehouse dedicated to model inference.

Question 9

An engineering team is configuring Databricks Asset Bundles (DABs) to automate the deployment of a Mosaic AI Agent application and its underlying serving endpoints. Which TWO configuration steps are required in the bundle configuration files?

Accepted Answer

Define a resources block specifying the model serving endpoints and app definitions within the bundle YAML configuration.. Databricks Asset Bundles rely on structured YAML configuration files to define project metadata, resources, and deployment targets. Properly declaring the workspace paths and resource types enables seamless synchronization and deployment of code and serverless endpoints using the Databricks CLI. This ensures reproducible and automated promotion across development, staging, and production environments.

Answer

Include a hardcoded personal access token directly inside the default target block to authenticate automated CI/CD pipelines.

Answer

Configure a manual notebook execution step that compiles the bundle files into a compiled Java archive before deployment.

Answer

Set the runtime environment to execute exclusively on legacy Hadoop MapReduce clusters managed outside the workspace.

Question 10

Refer to the exhibit.

Accepted Answer

The app.yaml configuration file is missing or contains invalid syntax required for provisioning the application runtime.. The deployment failure stems from a missing or improperly structured app.yaml file required by the Databricks Apps runtime environment. When deploying modern generative AI applications, the platform expects a clear manifest detailing the execution command and environment configuration to provision serverless resources successfully.

Answer

The database schema migration failed because the SQL warehouse was suspended during deployment.

Answer

The application source code exceeds the maximum file size limit imposed by the workspace storage volume.

Answer

The personal access token used for authentication has expired during the git synchronization phase.

Question 11

Refer to the exhibit. An engineer is attempting to deploy a model with a custom GPU instance type. The deployment fails. What is the most likely cause based on the provided snippets?

Accepted Answer

The compute_config format is not valid for Databricks Model Serving.. The deployment fails because the provided compute configuration is attempting to use a specific instance type that is either unavailable in the current region or not supported by the Model Serving service's serverless infrastructure. While MLflow logs the model correctly, the serving environment restricts hardware types to predefined, managed configurations. Choosing unsupported hardware leads to provisioning errors during the endpoint creation phase in Databricks.

Answer

The model signature is missing the input schema definition.

Answer

The pip_requirements list is incompatible with the serverless environment.

Answer

The model artifact_path must be a full S3 URI.

Question 12

What is the correct way to handle sensitive configuration values (like API keys) in a Databricks Asset Bundle to ensure they are not committed to version control?

Accepted Answer

Reference the secret using the ${secrets} lookup syntax. Best practices for sensitive data in DABs involve using environment variables that are injected at runtime or referring to secrets stored in Databricks Secret Scopes. By using variable references (e.g., '${secrets.scope.key}') in the 'databricks.yml' file, the bundle avoids exposing the actual secret value in the source code. This ensures that the code can be safely shared and versioned while keeping sensitive credentials protected within the Databricks environment.

Answer

Store the secret in a .env file and include it in the bundle

Answer

Use a comment in the YAML file to warn users not to share it

Answer

Encrypt the secret string directly inside the YAML file

Question 13

A GenAI engineering team has a Mosaic AI Agent application packaged as a Databricks Asset Bundle. They need the `databricks bundle deploy` command to target the production workspace and use production-specific values for a `catalog_name` variable, without editing the bundle files. Which approach should they use?

Accepted Answer

Define the variable in the `targets.prod` block of `databricks.yml` and run `databricks bundle deploy -t prod`.. Databricks Asset Bundles let you declare environment-specific values inside target blocks in `databricks.yml`. Placing `catalog_name: prod_catalog` under `targets.prod.variables` and deploying with `-t prod` applies the production value automatically and reproducibly, with no file edits required at deploy time. This is the documented way to manage per-environment variables for a Mosaic AI Agent application bundle.

Answer

Run `databricks bundle deploy -t prod --var="catalog_name=prod_catalog"`.

Answer

Run `databricks bundle deploy -t prod --var catalog_name=prod_catalog`.

Answer

Create a `databricks.yml` copy named `databricks.prod.yml` with the production values and run `databricks bundle deploy -f databricks.prod.yml`.

Question 14

An engineer is building a RAG application that must query a vector search index containing confidential financial documents. The application will be deployed as a Model Serving endpoint, and only authorized users should see retrieved content. Which combination of Databricks features should the engineer implement to enforce this?

Accepted Answer

Enable inference tables on the endpoint and configure row-level filters on the vector search index.. Row-level filters on the Vector Search index enforce that each query only retrieves documents the requesting user is permitted to see, which is essential for confidential content. Inference tables add request and response logging for auditing and monitoring, so combining these two features addresses both access control and observability in a production RAG deployment.

Answer

Store the documents in DBFS and apply a cluster-level init script that filters results before returning them.

Answer

Configure the endpoint with a service principal and set the vector search index to public read access.

Answer

Use a personal access token in the application code and rely on Unity Catalog table ACLs on the source documents.

Question 15

A GenAI engineering team at a retail company has built a RAG application using Databricks Mosaic AI Agent Framework. They package the agent with MLflow and deploy it to a Databricks Model Serving endpoint that is consumed directly by an internal web front end. The front end must send the end user's identity to the agent so that the agent's retriever can filter product documents by the user's department. Which mechanism should the team use to pass this per-request user identity to the deployed agent?

Accepted Answer

Have the front end pass the identity in the model serving request's Databricks options and read it in the agent via the MLflow context.. The agent must receive the end user's identity on every request so the retriever can filter documents by department. Databricks Model Serving forwards caller-provided options to the served model, where the MLflow context exposes them to the agent code. This keeps identity per-request instead of collapsing it into the endpoint's service principal, which is what the department filtering in the scenario requires.

Answer

Enable Unity Catalog row filters on the underlying document table and let the endpoint's service principal determine the filter.

Answer

Store the user's department in a Delta table and have the agent look it up using the endpoint's service principal credentials at inference time.

Answer

Create a separate model serving endpoint per department and route users by having the front end choose the endpoint URL.

Question 16

A GenAI engineer has built a RAG chain with LangChain and logged it to Unity Catalog using MLflow. The chain's retriever queries a Vector Search index that is rebuilt nightly. The engineer must deploy the chain to a Databricks Model Serving endpoint so that the deployed version always reads the latest index without redeploying the endpoint. Which approach should the engineer take?

Accepted Answer

Log the chain with mlflow.langchain.log_model and reference the Vector Search index by name in the model signature; the endpoint resolves the index at query time.. The key requirement is that retrieval must track an index that changes nightly without redeploying the endpoint. Separating the chain from the index and referencing the index by name at query time achieves this, because each inference request performs a fresh Vector Search lookup against the current index contents. Embedding data in artifacts or caching lookups both freeze or delay freshness.

Answer

Deploy the chain as a Databricks job that runs on a schedule and writes answers to a Delta table that the application reads.

Answer

Bake the index contents into the logged model artifacts so the endpoint serves a self-contained snapshot of the embeddings.

Answer

Register the chain in Unity Catalog and configure a serving endpoint that caches index lookups for 24 hours.

Question 17

A GenAI engineer is packaging a Retrieval-Augmented Generation application as a Databricks Asset Bundle and wants `databricks bundle deploy` to also create the underlying Vector Search index and the serving endpoint that hosts the agent. Which TWO bundle resource declarations should the engineer include? (Choose two.)

Accepted Answer

A `resources.serving_endpoints` block that defines the endpoint name, config, and the served model.. Databricks Asset Bundles support declarative resource types for the components of a RAG application. The `serving_endpoints` resource creates the Model Serving endpoint that hosts the agent, and the `vector_search_indexes` resource creates the Vector Search index that powers retrieval. Together they let `databricks bundle deploy` provision both the retrieval layer and the serving layer from a single version-controlled bundle.

Answer

A `resources.model_serving` block that wraps the serving endpoint under a model-specific resource type.

Answer

A `resources.mlflow_experiments` block that registers the agent's chain as an MLflow model.

Answer

A `resources.registered_models` block that points to the Unity Catalog model to be served.

Question 18

An engineer is packaging a Mosaic AI Agent application for deployment using Databricks Asset Bundles. The bundle must (1) deploy the agent code and configuration to a target workspace from a CI pipeline, and (2) register the resulting model so a serving endpoint can reference a specific version. Which TWO bundle elements must the engineer define to accomplish this? (Choose two.)

Accepted Answer

A `targets` block mapping environments such as dev and prod to workspace hosts and configuration.. Databricks Asset Bundles express deployable resources and deployment environments through two core blocks. The resources block declares the model and serving endpoint so `bundle deploy` creates or updates them and links the endpoint to a model version. The targets block maps dev and prod to distinct workspaces and configuration, enabling CI-driven promotion. Permission, sync, and artifact blocks support deployment but do not register the model or define the endpoint.

Answer

A `permissions` block that assigns CAN_MANAGE to the CI service principal on every bundle resource.

Answer

A `sync` block that uploads the local notebook and Python files to the workspace on each deploy.

Answer

An `artifacts` block that compiles the agent into a wheel for deployment.

Question 19

A company runs a customer-facing GenAI assistant on a Mosaic AI Model Serving endpoint. During a peak-traffic incident, the endpoint began returning HTTP 429 errors and some requests timed out. The team wants the endpoint to absorb bursts without failing requests while keeping idle cost low. Which configuration change should the engineer make?

Accepted Answer

Increase the endpoint's provisioned concurrency and enable scale-to-zero when traffic drops.. HTTP 429 responses from a serving endpoint signal that the concurrency ceiling was reached. Provisioned concurrency determines how many requests can be processed simultaneously, so raising it lets the endpoint absorb traffic bursts. Pairing that with scale-to-zero keeps replicas from running when idle, controlling cost. Timeout reduction, smaller GPUs, and payload logging do not increase request-handling capacity and do not stop the throttling.

Answer

Move the endpoint to a smaller GPU SKU so more replicas fit within the same budget.

Answer

Lower the endpoint's query timeout so slow requests fail faster and free capacity.

Answer

Enable payload logging to inference tables so the endpoint can queue requests more efficiently.

Question 20

An engineer has a Mosaic AI Agent application in a Databricks Asset Bundle. The bundle defines a model serving endpoint whose configuration references a secret scope for an external LLM provider API key. The engineer runs `databricks bundle deploy` targeting the production workspace and the deployment succeeds, but at runtime the agent receives 401 Unauthorized responses from the provider. The secret scope and key already exist in the production workspace with the same name and permissions as in development. Which action should the engineer take to resolve the runtime authentication failure?

Accepted Answer

Grant the serving endpoint's service principal the READ permission on the secret scope, and verify the endpoint's environment references the secret via `{{secrets/...}}` syntax.. Model Serving resolves secrets using the endpoint's service principal identity, which needs READ on the secret scope. The 401 indicates the endpoint either lacks permission or is sending a literal placeholder instead of the resolved secret value. Granting READ and ensuring the configuration uses the `{{secrets/scope/key}}` reference syntax restores authenticated calls to the external provider.

Answer

Restart the model serving endpoint with `databricks serving-endpoints update-config` to force the secret to be re-read at startup.

Answer

Store the API key as a plaintext environment variable in the bundle's serving endpoint configuration so the endpoint can read it directly.

Answer

Add the secret scope and key names to the bundle's target configuration, then redeploy with `databricks bundle deploy`.