Free ebookVector DBRAGHNSW

Vector databases for RAG

The production reference for vector databases in RAG: indexing choices, HNSW vs IVF, hybrid search wiring, scaling cost, and when to switch stores. Grounded in the agentic-rag ChromaDB notebook and the complex-RAG-guide.

Inside the ebook

What you get

  • Pick a vector store based on corpus size, QPS, and operational surface
  • Tune HNSW parameters without guessing (M, efConstruction, efSearch)
  • Wire hybrid search across vector and BM25 indexes cleanly
  • Estimate cost across Chroma, Qdrant, Pinecone, pgvector at realistic scale
  • Plan a migration without downtime by running dual-write

Inside

  • The five axes that decide your vector store
  • HNSW vs IVF vs flat, with numbers
  • Hybrid search wiring (vector + BM25 + RRF)
  • Metadata filters without killing recall
  • Cost and scaling math for four common stores
  • Migration playbook: dual-write + cutover
  • Ship checklist
Checking access…
Keep going

What to do next

Go all the way: accelerator

AI Engineer Career Accelerator Program

Ship a portfolio of production AI systems and defend every design choice in senior AI interviews.

See the program

Course deep-dive

Hybrid document search with Qdrant and Sentence Transformers

Hybrid retrieval with dense vectors, sparse keywords, RRF fusion, and cross-encoder reranking.

Open course