Design a Typeahead Search System

Design a typeahead search experience. As a user types, return useful suggestions with low perceived latency and explain how frontend request handling fits the backend retrieval design. ### Constraints & Assumptions The source names typeahead and notes possible short coding follow-ups, but gives no specific domain or algorithm. State candidate data, matching rules, personalization, and freshness requirements before sizing the system. ### Clarifying Questions Are suggestions queries, products, people, or documents? Is matching prefix-only or tolerant of spelling errors? What ranking and access-control rules apply? How much update delay is acceptable? ### What a Strong Answer Covers An API, retrieval/index choice, ranking, caching, request cancellation or stale-response protection, and evaluation of latency and usefulness. ### Follow-up Questions Why can debouncing alone still show stale results? What belongs in a cache key? How would you support a changing corpus or permission-restricted suggestions?