Evals And Observability easy
Your team edits the system prompt of a production chatbot about twice a week. Last month an edit quietly broke refund-policy answers. What process change best prevents this?
- A Run a full offline eval once per quarter and fix any regressions that the quarterly report happens to surface.
- B Make the offline eval a required CI check on every prompt or model change, blocking merges that regress.
- C Have each author spot-check three or four refund conversations manually before merging their prompt edits.
- D Monitor thumbs-down rates in production and roll back any prompt edit after the weekly rate rises noticeably.
Answer: B. Make the offline eval a required CI check on every prompt or model change, blocking merges that regress.
Evals are a gate on every change, not a periodic audit; the cost is a few minutes of CI time and judge tokens per merge. Watching thumbs-down rates seems sufficient, but online signals lag by days and users absorb the regression first. Production monitoring is still needed, as a backstop for failures the eval set does not cover.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- In one release, a RAG assistant got a new embedding model and a new chunking strategy, and the end-to-end judge score dropped 6 points. The team only tracks the end-to-end score and has spent a week tweaking the generation prompt with no recovery. What is the best next step?
- Over two months, your team iterated through 40 prompt versions, each tuned against the same 150-case golden set. The offline score rose from 72 to 91. Online task completion and retry rate have not changed. What is the most likely problem, and what is the fix?
- A prompt change produced an impressive answer in a live demo for leadership. The PM wants to ship it today. What should you do?
- You plan to replace human graders with an LLM judge for your answer-quality eval. Before trusting its scores in a CI gate, what should you do first?
- After you swap to a smaller model, p95 latency falls from 900 ms to 350 ms and the ops dashboard is all green. Leadership asks whether answer quality is still fine. What is the right response?
- Your CI gate fails any prompt change that drops the score 3 or more points on a 100-point rubric over 60 cases. Production generates at temperature 0.7. Running the eval twice on the same commit gives scores that differ by up to 4 points, and engineers have started rerunning failed gates until they pass. How do you fix the gate?