Evals And Observability easy
After switching to a smaller model, your dashboard shows p95 latency fell from 4s to 1.5s and cost per request halved. A manager calls it a clear win. What is missing?
- A Nothing; lower latency usually means the model is answering more directly, which users read as higher quality.
- B The p50 latency, since p95 is dominated by outliers and does not reflect what typical users experience.
- C A longer observation window for cost, since token prices fluctuate and halving may not hold next month.
- D A quality eval on the same traffic, since latency and cost say nothing about answer correctness.
Answer: D. A quality eval on the same traffic, since latency and cost say nothing about answer correctness.
Latency, cost, and quality are independent axes, so a smaller model can be faster and cheaper while answering worse. Treating speed as a proxy for quality is the tempting mistake. A dashboard needs a quality metric beside latency and cost. Latency alone decides the question only when quality has already been shown to be equal on your eval set.
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?