Michael Psenka
Research Scientist Intern (adv. Yann LeCun) at Meta
San Francisco Bay Area United States
Summary
👤
Senior🎓
Top SchoolMichael Psenka is a research-focused engineer and EECS PhD candidate at UC Berkeley with eight years of experience bridging mathematical theory and practical machine learning systems. Currently interning with Yann LeCun at Meta, he works on enabling reasoning and planning in real-world video world models that interface with action and control. His background includes provable advances in optimization on tensor manifolds, non-Euclidean multi-view reconstruction, and applied ML systems work—from denormalizing databases and search improvements to a workplace-analytics algorithm and a patent-driven product flow. He has taught and organized large undergraduate courses in discrete math and probability, demonstrating an ability to translate deep theory into clear instruction. Notably, his work consistently seeks to extend pretrained models beyond their initial capabilities through principled, mathematical approaches that scale to embodied and control-aware settings.
8 years of coding experience
1 year of employment as a software developer
Doctor of Philosophy - PhD, Electrical Engineering and Computer Science (EECS), Doctor of Philosophy - PhD, Electrical Engineering and Computer Science (EECS) at University of California, BerkeleyBachelor of Arts - BA, Mathematics, minors in Computer Science and Applied Math, Graduate, Bachelor of Arts - BA, Mathematics, minors in Computer Science and Applied Math, Graduate at Princeton University
Github Skills (9)
autoencoder9
matlab9
unsupervised-learning8
representation-learning8
geometry7
ai5
machine-learning5
deep-learning4
pytorch1
Programming languages (3)
Github contributions (5)
michael-psenka/michael-psenka.github.io
Jun 2021 - Apr 2026
Contributions:20 pushes in 4 years 10 months
michael-psenka/manifold-linearization
May 2023 - Jul 2024
Companion repository for the paper "Representation Learning via Manifold Flattening and Reconstruction"
Contributions:2 PRs, 33 pushes, 2 branches in 1 year 2 months
representation-learningaiautoencoderdeep-learninggeometry
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.