Author here. I've been using PostgreSQL for AI workloads in production for the past 3 years and wrote this book because I kept seeing teams spin up Pinecone, Weaviate, or Chroma when pgvector could handle their use case with less operational overhead.

The book covers:

- Vector similarity search with pgvector (HNSW vs IVFFlat, indexing strategies, hybrid search)

- Building RAG pipelines entirely inside PostgreSQL

- Recommendation systems using collaborative filtering in-database

- Feature engineering with pure SQL

- In-database ML with PostgresML

- Production patterns (monitoring, scaling, CDC with Debezium)

Everything runs on a single Docker Compose stack: PostgreSQL 17 + pgvector + TimescaleDB + Ollama. No GPU required, a laptop with 16 GB RAM is enough. Two tiers: $29 for PDF, $49 for PDF + EPUB + complete source code with Docker environment.

There's a free sample chapter on the landing page. Happy to answer any questions about PostgreSQL + AI.