How Vector Search Works: IVF and HNSW Explained

Vector search powers semantic systems like RAG and recommendation engines by finding approximate nearest neighbors. IVF clusters data into groups for fast lookup, while HNSW builds a multi-layer graph for efficient navigation. Both trade exactness for speed, using recall as a quality knob.