AI Engineering: Building Applications with Foundation Models by Chip Huyen

The discipline shifted. Where machine learning engineering meant training models, AI engineering means building products on models somebody else trained — a different job with different failure modes. Chip Huyen’s book is the first serious attempt to define that job rather than demonstrate a framework.

What the book covers

• What foundation models can and cannot do, and how training choices shape behaviour
• Evaluation: why it is the hardest part, and how to build evaluation pipelines you trust
• Prompt engineering as an engineering practice, including defensive prompting
• Retrieval-augmented generation and context construction
• Finetuning — when it beats prompting and retrieval, and when it does not
• Dataset engineering, data quality and synthetic data
• Inference optimisation: latency, throughput and cost
• Architecting the full application, agents, and feedback loops in production

Who it’s for

Engineers and technical leads shipping applications on top of LLMs, ML practitioners adapting to foundation models, and product people who need to understand the constraints they are designing against.

Edition details

• Author: Chip Huyen
• Publisher: O’Reilly Media
• Edition: 1st Edition
• Format: Paperback
• ISBN: 978-1098166304
• Language: English

Ships free within the United States from Shelfcove.