Fundamentals

Vector Database

A vector database stores and retrieves embeddings efficiently. Instead of traditional databases optimized for exact matches, vector databases use specialized algorithms (like HNSW or IVF) to quickly find similar vectors. They enable semantic search at scale, supporting millions of documents with sub-millisecond latency. Vector databases are essential infrastructure for RAG, recommendation systems, and semantic search applications. Popular options include Pinecone, Weaviate, Milvus, and Qdrant. They typically support metadata filtering, hybrid search, and various distance metrics.