Sebastian Raschka

Sebastian Raschka is a machine learning and AI researcher and LLM research engineer who focuses on the practical implementation and comparison of large language model architectures. His content includes building models from scratch, covering topics such as pretraining on unlabeled data, instruction fine-tuning, text generation with KV caching, and implementing reasoning models with verifiers. He also explores modern architectures like recurrent depth and looped transformers, as well as techniques such as Claude's text watermarking.

He is the author of the book "Build a Large Language Model From Scratch" and creates video tutorials that correspond to the implementation process. The channel provides a visual tour and comparison of different LLM building blocks and transformer alternatives, detailing the lessons learned from hands-on implementation.