Ingmar Kanitscheider
Research Scientist at OpenAI
Zürich Metropolitan Area Switzerland
Summary
🤩
Rockstar🎓
Top SchoolIngmar Kanitscheider is a research scientist at OpenAI with nine years of experience specializing in tuning large language models and optimizing GPT-4. Trained as a theoretical physicist (PhD) with advanced degrees from Cambridge and Amsterdam, he transitioned from a decade of computational neuroscience and Bayesian modeling into applied deep learning. His prior work produced SLAM-comparable navigation models and neural theories that revealed hippocampal-like coding and efficient coding explanations for Weber’s law. He combines rigorous mathematical training with hands-on hyperparameter and optimization expertise, making him adept at squeezing performance from large-scale models. Based in the Zürich area, he brings a rare blend of theoretical insight, experimental neuroscience grounding, and practical model engineering. Notably, his background means he approaches LLM tuning with a principled, inference-driven perspective rather than purely empirical trial-and-error.
9 years of coding experience
Master’s Degree, Theoretical and Mathematical Physics, Distinction, Master’s Degree, Theoretical and Mathematical Physics, Distinction at University of CambridgeDoctor of Philosophy (Ph.D.), Theoretical Physics (String theory), Doctor of Philosophy (Ph.D.), Theoretical Physics (String theory) at University of AmsterdamUndergraduate studies, Physics and Computer Science, Undergraduate studies, Physics and Computer Science at The University of Bonn
English, German, French, Spanish, Dutch, Russian
Github Skills (13)
reinforcement-learning7
python6
read-write4
simulation4
simulations4
deep-reinforcement-learning4
physics3
machine-learning3
numpy-arrays3
tensorflow3
mujoco2
physics-engine2
cpp2
Programming languages (4)
Github contributions (5)
ingkanit/multi-snake-RL
Mar 2018 - Mar 2018
Contributions:2 commits, 2 pushes, 1 branch in 1 day
minerllabs/minerl
Feb 2021 - May 2021
MineRL Competition for Sample Efficient Reinforcement Learning - Python Package
Contributions:5 commits, 6 pushes, 1 branch in 3 months
pythonminerlreinforcement-learningminerl-competitionlearning-python
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.