ML//RAG//hybrid search

Hybrid search is a retrieval method that combines vector similarity (semantic) with keyword matching (BM25), used because neither alone is sufficient: vectors miss exact terms, and keywords miss meaning. Reciprocal rank fusion merges the two ranked lists, and most production RAG systems use hybrid search.

Hybrid search is a retrieval method that combines vector similarity (semantic) with keyword matching (BM25), used because neither alone is sufficient: vectors miss exact terms, and keywords miss meaning. Reciprocal rank fusion merges the two ranked lists, and most production RAG systems use hybrid search.