MentoroAI
Back to Glossary

Vector Database

Trends & Concepts

Index for similarity search in embedding spaces.

Vector databases store embeddings and enable fast similarity searches.

  • Index structures: HNSW, IVF-PQ, DiskANN, etc.
  • Parameters: Dimensionality, distance metrics, recall/speed trade-offs.
  • Practice: Deduplication, metadata filtering, re-indexing, observability.