Design product search that supports free-text queries plus criteria such as price range, category, and nested subcategory. The catalog and request volume are large, filters can be combined, and results need deterministic pagination. ### Constraints & Assumptions - Product price and availability change more often than most descriptive fields. - A product may belong to more than one taxonomy node. - The design must explain indexing, query execution, ranking, and update propagation. - Use qualitative scale assumptions unless the interviewer supplies concrete numbers. ### Clarifying Questions to Ask - Are results keyword-ranked, filter-only, or a blend of relevance and business rules? - How fresh must price and availability be in search results? - Are facet counts required for every query, and do users need locale-specific prices? ### What a Strong Answer Covers - A source-of-truth catalog separated from a denormalized search index. - Exact and range fields for taxonomy and price, plus analyzed text fields for searchable content. - A query plan that applies filters, ranking, stable tie-breaks, and cursor pagination. - Incremental indexing with version checks, replay, reconciliation, and a strategy for stale price data. - Taxonomy evolution, multi-region serving, hot queries, cache boundaries, overload behavior, and quality metrics. ### Follow-up Questions - How would you change the design if price must be exact at checkout but search may be slightly stale? - How would you reparent a large subcategory without blocking search? - How would you detect that one filter combination silently returns incomplete results?