Ian Fox
Research Scientist at Meta
Shrewsbury, England, United States
Summary
🤩
Rockstar🎓
Top SchoolIan Fox is a research scientist at Meta with 10 years of experience specializing in ML-driven healthtech, currently developing computer vision and signal-processing algorithms for mobile cardiovascular monitoring. He previously applied reinforcement learning and offline evaluation to IG Reels ranking and notification scheduling and contributed to Facebook’s open-source ReAgent platform, refactoring DQN/TD3 trainers to PyTorch Lightning and adding CRR and evaluation fixes. Ian earned a PhD from the University of Michigan where he developed ML techniques for physiological time series and deep RL for an artificial pancreas, with papers at KDD, ICML, IJCAI and MIT Sloan Sports Analytics. Based in Shrewsbury with US training in math and CS, he bridges rigorous research and production engineering to push clinical-grade algorithms onto consumer devices.
11 years of coding experience
5 years of employment as a software developer
Master's degree, Computer Science, GPA 3.95, Master's degree, Computer Science, GPA 3.95 at University of MichiganBachelor's degree, Mathematics and Computer Science, GPA 3.99, Bachelor's degree, Mathematics and Computer Science, GPA 3.99 at University of Massachusetts Amherst
Github Skills (8)
pytorch10
machine-learning10
dqn10
pytorch-lightning10
tdd10
python10
reinforcement-learning10
unit-testing8
Programming languages (2)
Github contributions (5)
facebookresearch/ReAgent
Oct 2020 - Sep 2021
A platform for Reasoning systems (Reinforcement Learning, Contextual Bandits, etc.)
Role in this project:
ML Engineer
ML EngineerContributions:24 commits, 17 PRs, 1 branch in 11 months
Contributions summary:Ian's primary contribution focused on refactoring and converting existing TD3 and DQN trainers within the ReAgent framework to utilize PyTorch Lightning. They introduced new reporter classes and modified existing code to integrate with the new Lightning modules. Further contributions include adding a CRR trainer, modifying the model registration process, fixing bugs related to the Evaluator and reward boosts within the DQN and QR-DQN trainers, and adding unit tests.
reinforcement-learningcontextualbanditscontextual-banditsreinforcement
igfox/multi-output-glucose-forecasting
May 2018 - Dec 2021
Contributions:9 commits, 5 pushes, 1 branch in 3 years 6 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.