hybrid-search-implementation - Implement hybrid vector and keyword search
Combines vector similarity and keyword search with fusion or reranking for improved retrieval in RAG and search systems.
Tags
Updated: 2026-10-04Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Combine vector and keyword results
- Fuse ranked result lists
- Normalize and combine scores
- Configure PostgreSQL hybrid search
- Configure Elasticsearch hybrid search
- Rerank candidates with cross-encoders
Inputs
- Search query
- Query embedding
- Ranked result lists
- Document corpus
- Metadata filters
- Database connection pool
- Elasticsearch client
- Cross-encoder model
Outputs
- Fused search results
- Reranked search results
- Database schema and indexes
- Search index configuration
Requirements
- Python environment
- PostgreSQL with pgvector
- PostgreSQL full-text search
- Elasticsearch for Elasticsearch implementation
- Embedding model support
- Cross-encoder model support for reranking