Sebastian Raschka

ML researcher, author of Build a Large Language Model (From Scratch) and Build a Large Reasoning Model (From Scratch); writes about LLMs, agents, and deep learning.

Last updated: 2026-04-12

Overview

Sebastian Raschka is a machine learning researcher and educator best known for his accessible, from-scratch implementations of ML concepts. He writes the Ahead of AI newsletter on Substack covering LLMs, reasoning models, agents, and practical deep learning.

Key Works

• Build a Large Language Model (From Scratch) — step-by-step book implementing a GPT-style LLM from scratch
• Build a Large Reasoning Model (From Scratch) — follow-up covering reasoning model training; early access as of 2026
• Mini Coding Agent — minimal Python implementation of a coding agent illustrating the 6 harness components
• Ahead of AI newsletter — Substack covering LLMs, attention variants, agents

Key Contributions to Wiki

• Framing of LLM vs. reasoning model vs. agent distinction
• Six-component breakdown of coding agent harnesses — see coding-agent

Connections

• coding-agent — his article is the primary source for this page

Sources

• Components of A Coding Agent — added 2026-04-12