Is This Job For You?
A five-minute orientation before the real chapters: what an AI Search / RAG engineer actually does, whether it fits you, and how to get everything out of this book for free.
The job, in one sentence
You are the person who makes a language model answer truthfully about things it was never trained on — your company's docs, a file uploaded five minutes ago, an API shipped last week — by finding the right few hundred tokens and putting them in front of the model at question time. Plugging in a model is easy now; making it stop confidently making things up is the hard, in-demand part, and that's your job.
Do you need a PhD or years of ML? No.
What you need is comfort reading and writing basic Python, curiosity about numbers (you'll meet a little retrieval math — recall@k, cosine similarity — all explained from scratch), and the instinct to ask "did we actually retrieve the answer?" instead of "the demo looked fine." If that sounds like you, you can do this job.
Why it's a great first role in AI
Almost every company shipping an LLM feature needs retrieval — "chat with your docs," support assistants, internal search — and very few engineers specialize in it. It's a way into AI that doesn't require training models: you work on top of them, wiring the retrieval that makes them useful. And because the discipline is only a few years old, there are no twenty-year veterans to compete with.
How to use this book
Each chapter is a short worked read → a real case study with numbers → a hands-on Try-It lab → a worked solution → a quiz that needs reasoning, not recall. A single company, Acme Vault, is built up chapter by chapter so you see the whole pipeline come together. Do the labs. Every one runs free — no API key, no credit card, no spend — so there's no excuse to just read. By the final chapter you'll have built a small but real RAG pipeline, with an eval harness that proves it works, that you can keep on your laptop and show in interviews. That portfolio piece is worth more than any certificate.