Google Cloud · Agentic Architect
Continuous evaluation and golden datasets
Agentic Architect · Lesson 12 of 16 · 1:40:57
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 Google Cloud Professional Agentic Architect exam, on the official exam guide. In this lesson: success criteria and thresholds, evaluation in the CI/CD pipeline, golden datasets for RAG, and online monitoring. Includes three original practice questions.
Preview the topics
- A change to an instruction must never reach production if the agent’s tool selection score falls below its threshold. What should the team set up?
- Cymbal Air’s baggage policy answers are often wrong, and the golden dataset shows that the expected policy document is almost never retrieved. What should the team improve first?
- Over several weeks, the quality of the agent’s answers slowly drops in production, although nothing was deployed. How should the team catch this earlier next time?