Glossary
Vector Database
A vector database stores embeddings — arrays of numbers that represent the meaning of text, images, or other content — and retrieves items by similarity: given a query vector, find the stored vectors nearest to it. That nearest-neighbour search is what powers semantic search (“find documents about refund policy” without matching the words) and retrieval-augmented generation (RAG), where an AI model is fed the most relevant context before answering.
Architecturally it’s a new kind of index rather than a replacement for analytical storage: production systems pair it with the warehouse or lakehouse, where the governed source data lives, and treat embeddings as derived data — rebuilt when models change, governed like any other pipeline output.