CoursesFree Quizzes
DatabricksGovernance

Model Serving & Auditability — Databricks GenAI Associate Practice Question

A representative Databricks Certified Generative AI Engineer Associate (Databricks GenAI Associate) exam question on Model Serving & Auditability. Work through it below, then read why each option is right or wrong.

This question is part of the free Databricks Certified Generative AI Engineer Associate practice test.

Short answer

The correct answer is B. Configure Mosaic AI Model Serving to log all requests and responses to a Delta table, which can be queried and monitored for compliance audits..

Mosaic AI Model Serving provides native logging of requests and responses to Delta tables, enabling comprehensive audit trails. Delta tables integrate with Lakehouse Monitoring and can be queried for compliance verification. Logging to external databases introduces complexity and breaks the unified governance model. MLflow logging during training does not capture serving-time predictions. Workspace audit logs track administrative actions but do not capture model prediction details.

The Question

Your team is deploying a Mosaic AI Model Serving endpoint that generates product recommendations. Your organization's responsible AI policy requires that all model predictions be logged with their input features, output predictions, and a timestamp. Which approach best implements this requirement?

AUse MLflow's built-in logging to capture predictions only during model training, not during serving.
BConfigure Mosaic AI Model Serving to log all requests and responses to a Delta table, which can be queried and monitored for compliance audits.Correct
CHave the client application log predictions to a separate database outside Databricks.
DRely on Databricks workspace audit logs, which automatically capture all model predictions without additional configuration.

Why B is correct

Mosaic AI Model Serving provides native logging of requests and responses to Delta tables, enabling comprehensive audit trails. Delta tables integrate with Lakehouse Monitoring and can be queried for compliance verification. Logging to external databases introduces complexity and breaks the unified governance model. MLflow logging during training does not capture serving-time predictions. Workspace audit logs track administrative actions but do not capture model prediction details.

Why the other options are wrong

Option A: Use MLflow's built-in logging to capture predictions only during model training, not during serving.

Option A does not satisfy the requirement in the scenario. Review the explanation above: the correct choice (B) is the only one that fully meets every constraint stated in the question.

Option C: Have the client application log predictions to a separate database outside Databricks.

Option C does not satisfy the requirement in the scenario. Review the explanation above: the correct choice (B) is the only one that fully meets every constraint stated in the question.

Option D: Rely on Databricks workspace audit logs, which automatically capture all model predictions without additional configuration.

Option D does not satisfy the requirement in the scenario. Review the explanation above: the correct choice (B) is the only one that fully meets every constraint stated in the question.

Key idea: Model Serving & Auditability

Mosaic AI Model Serving provides native logging of requests and responses to Delta tables, enabling comprehensive audit trails. Delta tables integrate with Lakehouse Monitoring and can be queried for compliance verification. Logging to external databases introduces complexity and breaks the unified governance model. MLflow logging during training does not capture serving-time predictions. Workspace audit logs track administrative actions but do not capture model prediction details. On the Databricks GenAI Associate exam, questions in the "Governance" domain test whether you can map a scenario's constraints to the right choice. Read the requirement carefully, eliminate options that violate any single constraint, and pick the one that satisfies all of them with the least operational overhead.

Ready to see how you'd score?

Take the free practice quiz and find out which Databricks Certified Generative AI Engineer Associate domains you need to focus on. No signup required.

Take the free quizSee all 220+ practice questions

Practice 5 similar questions

Same cert, same or adjacent domain. Use these after reviewing the explanation.

1RAG & Vector Search — Databricks GenAI Associate Practice QuestionDesign Applications2Document Ingestion & Data Preparation — Databricks GenAI Associate Practice QuestionData Preparation3Agent Development & Tool Design — Databricks GenAI Associate Practice QuestionApplication Development4Agent Framework & Packaging — Databricks GenAI Associate Practice QuestionAssembling and Deploying Applications5Mlflow & Retrieval Evaluation — Databricks GenAI Associate Practice QuestionEvaluation and Monitoring

Related Databricks Certified Generative AI Engineer Associate Practice Questions

RAG & Vector Search — Databricks GenAI Associate Practice Question→Document Ingestion & Data Preparation — Databricks GenAI Associate Practice Question→Agent Development & Tool Design — Databricks GenAI Associate Practice Question→

Official sources used

Verify program and exam details at the source before scheduling.

docs.databricks.comEN: Agentsdocs.databricks.comMlflow3: GenAI