Databricks · Databricks Generative AI Engineer Associate
MLflow Tracing, Agents, Genie
Databricks Generative AI Engineer Associate · Lesson 9 of 17 · 1:00:58
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the Databricks Certified Generative AI Engineer Associate exam. In this lesson: MLflow traces with one span per step (autolog or decorator), building and deploying agents (ResponsesAgent, the legacy agents.deploy path, and the Agent Bricks CLI), a Genie Agent used as a tool, the permissions it needs, and the same scorers on an evaluation set and on sampled production traces. Includes three original practice questions.
Preview the topics
- Copperline’s agent gave one customer a wrong answer. The team needs to see which retrieval step and which tool call produced it, with their inputs and outputs. What should the team open?
- Copperline’s sales agent calls the order Genie Agent as a tool. The service principal that runs the agent gets an access error on every call. Which permissions does it need, at least?
- Copperline’s assistant is now in production. The team wants to check the safety of real answers every day, without labeling them, and without scoring every single request. What should the team set up?