Erik Jenner
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolErik Jenner is a research-focused ML engineer and PhD student in AI based in Berkeley, dedicated to reducing existential risk from advanced AI through technical research. With 11 years of experience spanning research internships and deep technical contributions, he now works at Google DeepMind on AGI safety and alignment. His work blends rigorous theory—papers on equivariant PDEs and interpretable reward models—with practical engineering, such as refactoring reward learning code in the widely used imitation PyTorch repository. He has a strong academic foundation (BS in Physics, MS in AI, and current PhD at UC Berkeley) and a track record of publishing and collaborating with leaders in the field. Notably, he focuses on making learned reward models more interpretable and robust, translating complex research into reusable open-source components.
11 years of coding experience
Doctor of Philosophy - PhD, Artificial Intelligence, Doctor of Philosophy - PhD, Artificial Intelligence at University of California, BerkeleyAbitur, 1.0, Abitur, 1.0 at Karls-Gymnasium StuttgartBachelor of Science - BS, Physics, 1.0, Bachelor of Science - BS, Physics, 1.0 at Ruprecht-Karls-Universität HeidelbergMaster's degree, Artificial Intelligence, Master's degree, Artificial Intelligence at University of Amsterdam
German, English, Latin, Greek
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (7)
imitation-learning10
pytorch10
machine-learning10
python10
reinforcement-learning10
gymnasium9
testing8
Programming languages (4)
Github contributions (5)
HumanCompatibleAI/imitation
Jul 2021 - Jan 2022
Clean PyTorch implementations of imitation and reward learning algorithms
Role in this project:
ML Engineer
ML EngineerContributions:74 reviews, 100 commits, 11 PRs in 5 months
Contributions summary:Erik primarily focused on refactoring and improving reward learning algorithms within the `imitation` repository. Contributions include refactoring RewardNet code, making it independent from AIRL, and implementing preference comparisons. The user also made adjustments to trajectory generation and implemented fixes for the adversarial trainer, demonstrating a focus on improving core components of the imitation learning framework.
pytorchimplementationsreinforcement-learningcleanmachine-learning
isolani-chess/web-client
Jun 2016 - Mar 2023
Contributions:2 PRs, 16 pushes, 4 branches in 6 years 9 months
web-client
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.