Hybrid Retrieval overview

Hybrid Retrieval as a cross-cutting topic — combining dense vector similarity with sparse keyword matching for better retrieval quality. Covers reciprocal rank fusion (RRF) for merging ranked lists, linear score combination with tunable weights, Weaviate's built-in hybrid search, Pinecone's sparse-dense vectors, and the empirical finding that hybrid consistently outperforms either method alone on diverse query types. The key design decision is the fusion strategy and how to weight semantic vs lexical signals.