Hybrid Search Blueprint Series: Semantic Boosting

Semantic Boosting is a hybrid search technique that combines vector search with full-text lexical search to improve results. It first executes a vector query to retrieve semantically similar candidates, then incorporates vector scores into a final lexical search. This approach forces the lexical engine to include and boost relevant documents, producing a refined result set. It elegantly solves complex query intents and can be applied with standard text search features. To implement, use a two-shot retrieval approach with a vector query followed by a lexical search.