Summary
Maxwill Lin is an experienced machine learning engineer and research-minded generalist with 11 years of experience, currently building foundation models for recommendation and advancing coding-agent behavior at Vmax. He has driven production improvements at Meta’s Recommendation System—training billion-parameter autoregressive models, accelerating evaluation pipelines from hours to minutes, and leading offline RL dataset design for long-term user value. His background spans applied research in cryptography, GNNs for NLI, and high-frequency quantitative work, giving him a rare blend of ML, systems, and math skills. Notable for rapid impact as a new hire (#1 code contributor at Meta) and for experimenting with LLM self-play, Recursive Eval, and Automated Environment Design, he combines hands-on engineering with research rigor. Based in Sunnyvale, he pairs top-ranked competitive problem-solving with a graduate CS degree from Georgia Tech and a track record of turning research ideas into production-ready systems.
11 years of coding experience