Build a Reasoning Model (from Scratch)

"Build a Reasoning Model (From Scratch)" provides a practical, step-by-step guide to understanding and constructing reasoning-oriented large language models, starting from a basic pre-trained model and progressing through methods like inference-time improvements and training-based techniques. The book emphasizes implementing core methods firsthand, illustrating how concepts such as self-refinement and reinforcement learning work, along with their trade-offs and failure modes. Designed for those familiar with Python and machine learning, it offers a deep, hands-on exploration of modern reasoning approaches in language models, all organized into a logically flowing, accessible format.