December 7–11

The LLM bootcamp is back — five days of hands-on training to build, deploy and secure LLM-powered applications.

Learn more →
Chris McCormick

Chris McCormick

AI Researcher and Educator, InnerWorkings AI

Chris McCormick is an independent AI researcher and educator focused on the inner workings of transformer models, especially the architectures and mechanisms that let them build meaning. His research series Patterns and Messages reframes attention using merged matrices to explore low-rank structure and efficiency trade-offs, and his write-ups and videos on mccormickml.com walk through topics such as Multihead Latent Attention, per-head output projections and FlashAttention’s variable-length mode. He has also ported GPT-2 speedruns to run on a single GPU, making pre-training experiments accessible on a modest budget. A long-time author of AI tutorials and code, Chris is a member of The RAG Pack, the group of AI educators teaching reinforcement learning for LLMs and Introduction to Agents, and led a hands-on GRPO training workshop at the Agentic AI Conference in April 2026. He studied at Stanford University.