AI Engineering: Building Applications with Foundation Models

Book Review: 4/5 Impact On Me (Book By Chip Huyen)

Read more about the book here.

I rate the impact of this book 4 out of 5 on me. It gave me a lot of relevant information for building an AI product, but much of the material was either not yet actionable or already reflected in our product.

I read the book while building an AI Copilot as a Product Manager. The book was valuable for things including better explanations for when to choose RAG, context engineering best practices, and evaluating AI copilot generation quality. The book helped with prompt iteration and AI evaluations.

Foundational model vs fine-tuning vs context engineering

I got a clearer sense of when to build or fine-tune a foundational model versus when context engineering is enough. In most cases context engineering is sufficient; the book helped me see where the line sits. In most cases, most companies should focus on context engineering, not fine-tuning.

Relevance to my work

I had worked around fine-tuning in Flex AI; parts of the book applied there. Other parts applied to the n8n AI Copilot. The book was useful in both contexts.

Overall

I skipped to the sections relevant to me and re-read where I expected more insight but was not getting it. Many sections reinforced what I already knew, so not everything yielded new insight. A lot of alpha or early best practices are still in labs or only show up in customer research, so I had to figure out next steps beyond what the book and our product already covered.