Kelvin Xu
Research Scientist at Meta
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolKelvin Xu is a research scientist based in Berkeley with 12 years of experience focused on large language models, the science of scaling, and learning from human interaction. After a PhD in Machine Learning from UC Berkeley and an MSc from Mila, he spent roughly four years at Google Brain/DeepMind working on pretraining (including Gemini 1.0–3.1) and has transitioned to Meta's TBD Labs to continue pretraining work. His background blends deep academic rigor—doctoral research and publications in robotics, meta-learning, and RL—with hands-on system-scale model development in industry. Notably, he bridges theory and practice, moving from robot manipulation algorithms to large-scale language model pretraining pipelines, and brings a cross-disciplinary perspective shaped by engineering training at the University of Toronto.
12 years of coding experience
9 years of employment as a software developer
Doctor of Philosophy - PhD, Machine Learning, Doctor of Philosophy - PhD, Machine Learning at University of California, BerkeleyMSc, Computer Science, MSc, Computer Science at Mila - Quebec Artificial Intelligence InstituteBASc, Engineering Science, BASc, Engineering Science at University of Toronto
Github Skills (34)
python10
data-science10
expressions10
machine-learning10
mathematical10
ml10
pymc10
deep-learning10
tensorflow10
deep-neural-networks10
neural-network10
mathematical-expressions10
theano10
math8
nvidia-docker8
Programming languages (2)
Github contributions (5)
kelvinxu/Theano
Mar 2015 - Apr 2016
Contributions:71 pushes, 7 branches, 1 comment in 1 year 1 month
matharraypythonmeta-programminggpu
kelvinxu/kelvinxu.github.io
Jan 2018 - Dec 2019
Contributions:12 commits, 199 pushes, 1 branch in 1 year 11 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.