Sebastian Raschka | Build a Reasoning Model (From Scratch) - Author Discussion

We just finished reading Build a Reasoning Model (From Scratch) for August’s AI Book Club, and for this special session we’ll be joined by Sebastian Raschka for a live discussion about the book.
Rather than a presentation, this will be a casual conversation about the ideas behind the book, what went into writing it, and the rapidly evolving world of reasoning models.

We’ll talk about topics like inference-time scaling, self-refinement, reinforcement learning, GRPO, distillation, and what Sebastian has learned from building these techniques from scratch.
Bring your questions! We’ll leave plenty of room for audience discussion, and you’re welcome to join even if you didn’t finish the book (or haven’t read it at all).

## About the Book

Build a Reasoning Model (From Scratch) Author: Sebastian Raschka Published: July 2026
The book is a practical guide to understanding how modern reasoning-oriented LLMs work by building their core techniques step by step.
It starts with a pretrained language model and moves through text generation and evaluation, inference-time scaling and self-refinement, reinforcement learning with verifiable rewards, GRPO improvements, and finally distillation into smaller reasoning models.

Chapters 1 Understanding reasoning models
2 Generating text with a pretrained LLM
3 Evaluating reasoning models
4 Improving reasoning with inference-time scaling
5 Inference-time scaling via self-refinement
6 Training reasoning models with reinforcement learning
7 Improving GRPO for reinforcement learning
8 Distilling reasoning models for efficient reasoning

Get the book: Manning: Promo code AIBookClub should give you 45% off manning.com/books/build-a-reasoning-model-from-scratch

O’Reilly: learning.oreilly.com/library/view/build-a-reasoning/9781633434677/

Want to keep discussing the book and other AI engineering topics? Join the Flyte MLOps Slack group.