This repository provides resources for the upcoming book "AI Engineering," aimed at technical professionals building AI applications with foundation models. It offers a framework for adapting LLMs and LMMs to real-world problems, covering the end-to-end process from evaluation to deployment, and serves as a companion to traditional ML system design.

How It Works

The book focuses on the fundamentals of adapting foundation models, distinguishing AI engineering from traditional ML engineering. It addresses practical challenges like hallucinations, prompt engineering, Retrieval-Augmented Generation (RAG), agents, and fine-tuning, offering a systematic approach to building and optimizing AI applications. The content is illustrated with case studies and expert reviews, emphasizing timeless principles over rapidly changing tools.

Quick Start & Requirements

This repository contains resources related to a published book. There are no direct installation or execution commands. The book itself is available for purchase on Amazon, O'Reilly, and Kindle.

Highlighted Details

• Companion to "Designing Machine Learning Systems" (DMLS), focusing on foundation models vs. traditional ML models.
• Covers end-to-end AI application development, including evaluation, prompt engineering, RAG, agents, and fine-tuning.
• Emphasizes foundational knowledge over specific tools and APIs.
• Includes case studies and expert reviews from industry leaders.

Maintenance & Community

The repository is a work in progress, with resources to be updated. The author, Chip Huyen, is active in the AI community. Discussions about the book can be found on Twitter at @aisysbooks.

Licensing & Compatibility

The repository itself does not specify a license. The book is commercially published by O'Reilly Media.

Limitations & Caveats

The repository is marked as "[WIP]" (Work In Progress), indicating that content may be incomplete or subject to change. The book is described as technical and may dive deep into certain topics, with guidance provided for readers who prefer to skip more technical sections.