Sebastian Raschka's from-scratch video course on building reasoning models (magazine.sebastianraschka.com)

Sebastian Raschka has launched an extensive video course titled "Build a Reasoning Model (From Scratch)," which explores the intricacies of reasoning models in artificial intelligence. This course, consisting of eight coding videos totaling 11 hours, guides learners through the implementation of reasoning methods using a pre-trained Qwen3 model in Python and PyTorch. The content covers essential topics, including inference scaling, reinforcement learning with verifiable rewards (RLVR), and how reasoning models differ from conventional large language models (LLMs). The course serves as a sequel to his prior "Coding LLMs from the Ground Up," though it can be accessed independently. This initiative is particularly significant for the AI/ML community as it deepens the understanding of how reasoning models enhance LLM capabilities, enabling them to tackle more complex tasks through improved inference strategies and learning mechanisms. The course emphasizes building models from scratch, providing hands-on experience that helps demystify advanced concepts. As reasoning models become an integral part of modern AI agents—serving as the "engine" for sophisticated applications—the course equips practitioners with valuable skills to create and refine these powerful tools, fostering innovation in AI-driven problem-solving.