U2xAI
…
Quality, Cost & Token Economics

Latency budgets and UX tradeoffs Interview Questions

25 questions. All 25 carry a written answer.

  1. #1What is a latency budget and how would you allocate one across a RAG pipeline?ConceptIntermediate
  2. #2Why do you set latency targets at the 95th percentile rather than the mean?CalculationIntermediate
  3. #3Describe how streaming changes perceived latency without changing actual latency.ConceptFoundational
  4. #4At what point does latency stop mattering and quality take over?ConceptAdvanced
  5. #5How would you decide between a fast weak model and a slow strong one for autocomplete?CaseIntermediate
  6. #6Explain the UX options available when a response will take 30 seconds.CaseIntermediate
  7. #7Describe the latency requirements for a voice interface and why they are stricter.ConceptAdvanced
  8. #8How do reasoning models complicate a latency budget?ConceptAdvanced
  9. #9What is time-to-first-token and why is it the metric users actually feel?ConceptIntermediate
  10. #10Design the loading experience for an agent that runs for two minutes.CaseAdvanced
  11. #11How do you handle latency variance rather than average latency?ConceptAdvanced
  12. #12Explain when to move work to the background and notify the user later.CaseIntermediate
  13. #13What is the latency cost of adding a reranking step, and when is it worth it?CalculationAdvanced
  14. #14Describe how parallelizing tool calls affects an agent's latency profile.ConceptAdvanced
  15. #15How would you set a timeout, and what should happen when it fires?CaseIntermediate
  16. #16Explain the relationship between latency and cost in model selection.ConceptIntermediate
  17. #17Critique a design that shows a spinner for an unbounded amount of time.Artifact critiqueFoundational
  18. #18How does latency tolerance differ between a consumer and an enterprise workflow?ConceptIntermediate
  19. #19What UX patterns make waiting feel productive rather than broken?ConceptIntermediate
  20. #20Describe how you would measure whether latency is costing you conversions.CalculationAdvanced
  21. #21How do you budget latency when a third-party API's response time is outside your control?CaseAdvanced
  22. #22Explain the tradeoff between showing a fast partial answer and a slow complete one.CaseAdvanced
  23. #23What is the right latency target for an AI search feature, and how did you derive it?CalculationAdvanced
  24. #24How would you communicate a latency regression to users?CaseIntermediate
  25. #25Set a latency budget live for a feature I describe.InterviewAdvanced