Sourangshu Pal

Writing about production AI systems — RAG pipelines, LLMs, vector search, and the engineering that matters in practice.

A practical walkthrough of Retrieval-Augmented Generation — from chunking documents to serving answers — with working Python code.

The concepts under the hood you need before building production AI systems — tokens, context windows, temperature, and the things nobody tells you.

What vector databases actually do, how approximate nearest neighbor search works, and how to choose and configure one for your AI application.