Are you curious about how ChatGPT, LLMs, and RAG systems actually work under the hood?

Do you want to understand tokenization, embeddings, vectors, chunking, and retrieval in a way that’s clear, simple, and practical?

This book is your step-by-step guide to building and understanding Retrieval-Augmented Generation (RAG) — the backbone of modern AI assistants.

✅ What is RAG? – High-level overview with analogies that make sense.
✅ Why Chunking Matters – How to split text into the right sizes, with examples.
✅ Overlap Explained – Why overlapping chunks prevents context loss.
✅ RAG Flowchart – A clear pipeline from question → embedding → retrieval → answer.
✅ Model vs Embedding Model – What they are, how they differ, why both are needed.
✅ K in RAG – What Top-K retrieval means and how to tune it.
✅ Chunk Size Sweet Spot – Why ~500 tokens works best (Goldilocks rule).
✅ Token Math – How English words map to tokens, 5 vs 11 char average.
✅ Dimensions – What vector dimensions are, with examples from OpenAI & BGE models.
✅ Parameters – How “7B parameters” are calculated from tokens × dimensions.
✅ FAISS Explained – An open-source vector DB for blazing-fast similarity search.

Gowtham SB