Hello Reader,

A few weeks ago, DeepSeek made headlines by being the most powerful open-source reasoning model. I made a video recently to test its performance against OpenAI's o1 model. Spoiler: It's pretty good, but not without flaws!

But there's another important implication of DeepSeek:

Open-source models are making it easier (and cheaper) than ever to build useful tools—like study buddies, workflow automators, or even your own AI twin.

Andrew Ng believes that even though a lot of attention and hype on the AI foundation models, most of the opportunities will be in building AI applications.

That said, while anyone can slap together a basic AI app, the real impact lies in building robust solutions that solve real problems.

In today's email, I want to share 3 books that have helped me sharpen my understanding of modern AI models and applications.

It’s hard for books to keep up with the lightning-speed developments in AI today. But for me, books are a great way to structure my learning and cut through the hype (and avoid the social media FOMO 🥲).

👉 If you prefer watching the video breakdown instead, here it is:

​

1. Build a Large Language Model (From Scratch) by Sebastian Raschka

If you love understanding technology at a deeper level, this book breaks down the components of GPT-like models with clear visuals and examples. It’s a great way to grasp how LLMs operate.

📚 This book:

• Offers a peek inside the “AI black box” and a step-by-step guide to building your own large language model (LLM).
• Guides you to code every part of an LLM with the PyTorch library (and understand the underlying components too).
• Covers everything from pretraining and fine-tuning an LLM for different purposes.
• Best for: Developers who crave deep technical understanding and those who love tinkering. You do need some Python experience and basic math skills.

Git repo: https://github.com/rasbt/LLMs-from-scratch​

2. AI Engineering by Chip Huyen

This book is a comprehensive guide to AI engineering and building AI applications.

It discusses the rise of AI Engineering as a discipline in today’s tech landscape.

📚 Moreover, this book:

• Provides a broad overview of using AI foundation models—like large language and multimodal models - to build real-world applications.
• Provides a useful guide on what to consider before building an AI application (e.g. evaluating use case, set expectations, plan milestones and maintenance).
• Explains key concepts like prompt engineering, retrieval-augmented generation (RAG), AI agents, and fine-tuning.
• Shows you different approaches to evaluating AI models and systems.
• Best for: Teams integrating AI into real-world applications.

Git repo: https://github.com/chiphuyen/aie-book​

3. LLM Engineer’s Handbook by Paul Iusztin & Maxime Labonne

If you’re serious about building and deploying LLMs, this book is a full-blown practical guide.

📚 This book:

• Uses a running example of the LLM Twin project that you can adapt to your own use case.
• Shows you how to implement everything from data collection & engineering, to retrieval-augmented generation (RAG), fine-tuning, inference optimization, deployment and LLMOps.
• The book also introduces some popular tools like Hugging Face, ZenML, Docker, and MongoDB—stuff you might use for deploying AI in production.
• Best for: Engineers ready to ship LLM apps fast.

Git repo: https://github.com/PacktPublishing/LLM-Engineers-Handbook​

What AI books or tools are you loving lately? Hit reply—I’d love to hear!

Cheers,
Thu

​

Work with me:

If you want to dive deeper with Python and build real-world AI projects this year, check out this program👇. You'll join a community of 200+ learners who are building their portfolios while supporting each other along the way.