LLM Engineer's Handbook: Master the art of engineering large language models from concept to production

Took me by surprise. The plan was to get a step by step guide of building LLM Twin but I got more — a full package of the whole ecosystem of building any AI systems.

This book lacks depth and clarity, offering simplistic examples with little real-world value. Over-reliance on vendor services, AWS-centric content, and questionable originality of the text —possibly lifted from online sources—make it feel rushed and unprofessional. Unique phrases and sections of text in this book can be searched online to verify their true origins. Not suitable for those seeking a deeper understanding of LLM building and deployment. There are much better books available by recognised experts in AI.

I really, really, really wanted to like this book. It's a massive disappointment though. I mean ignore the fact that none of the code from the book works right now because of deprecated and incompatible packages - that's something that each Python enthusiast needs to learn to live with. The main problem with this book is that contrary to it's title, it's not a journey "from concept to production". It does the complete opposite - it takes a completely productionised repo and dissects it, but in a way that's completely unusable for you if you wanted to code a long from scratch. This is not a book for developers, it's a book for archaeologists. It's got a few nuggets, but because you cannot really follow it along, it's purely theoretical, you will learn more by asking an LLM to guide you through it step by step and create this from scratch (ironic, isn't it)