Benjamin Scellier
Principal Research Scientist at Rain
Zurich, Zurich, Switzerland
Summary
👤
Senior🎓
Top SchoolBenjamin Scellier is a Principal Research Scientist in Zurich with a decade of experience building physics-informed machine learning algorithms for energy-efficient, analog ML processors. He specializes in Equilibrium Propagation and other energy-based learning methods that aim to replace GPU-heavy training with processors that merge memory and computation and exploit analog physics. His work spans theory, algorithm design, and software: he maintains an open-source framework for simulating analog systems compatible with EP, bridging simulation to hardware. Past roles include research residencies at Google and X and a postdoc at ETH Zürich, reflecting a trajectory from deep-learning foundations to hardware-aware algorithmics. Trained at Mila (PhD in Deep Learning) with strong quantitative roots from École Polytechnique and NUS, he combines mathematical rigor and experimental pragmatism to tackle orders-of-magnitude improvements in AI energy efficiency.
10 years of coding experience
3 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Deep Learning, Doctor of Philosophy (Ph.D.), Deep Learning at Mila - Quebec Artificial Intelligence InstituteMaster’s Degree, Applied Mathematics, Master’s Degree, Applied Mathematics at Ecole polytechniqueMaster’s Degree, Statistics, Master’s Degree, Statistics at National University of Singapore
French, English, German, Chinese
Github Skills (7)
pymc10
mathematical10
mathematical-expressions10
theano10
expressions10
python8
math7
Programming languages (1)
Github contributions (5)
bscellier/Towards-a-Biologically-Plausible-Backprop
Mar 2016 - Jan 2020
Contributions:11 commits, 12 pushes, 1 branch in 3 years 10 months
bscellier/Theano
Oct 2016 - Feb 2017
Contributions:13 pushes, 8 branches in 3 months
python-librarymulti-dimensionalpythonevaluatemathematical-expressions
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.