Certainly a gap in the current literature and good coverage of patterns overall, however the book although touching the right points looks a bit outdated at release and limited in depth e.g. seems to be missing key patterns like reasoning models, telemetry in observability, RL in LLM context, no graph RAG, multi-modality (book is named Generative AI not LLM's)...The curse of writing a book in such a fast developing field.
ps - I like Chip Huyen's books and AI Engineering seems to be hugely successful. However I think she didn't do good for the community when she started writing books in stateless long tweet thread format which this book mentions at the beginning suggesting the book is following Chip's style...which causes a lack of overall informational unity and integrity of the content and how things are related, data backed context on why x is better than y etc...
Note: This review was written on the pre-release version on O'Reilly for Public Libraries
An outstanding book that goes over multiple design patterns for deploying generative AI applications. Is very well-researched and generalizes many cases that I already knew about, but hadn't thought about applying it in certain cases. A lot more general than their last book on Machine Learning Design Patterns, a very good read and very ambitious given how quickly this domain is changing.