Piotr Sokol
Postdoctoral Researcher at Champalimaud Foundation
Oeiras, Lisbon, Portugal
Summary
👤
Senior🎓
Top SchoolPiotr Sokol is a postdoctoral researcher in Oeiras, Lisbon, blending eight years of experience across probabilistic machine learning, neural networks and computational neuroscience to probe learning in biological and artificial systems. With a PhD in Computational Neuroscience and Machine Learning from Stony Brook University and prior work at the Netherlands Institute for Neuroscience, he bridges detailed biophysical modeling and large-scale neural networks. Piotr contributes to open-source scientific ML tooling—improving neural ODE adjoint methods and solver callbacks in the notable DiffEqFlux.jl project—bringing rigorous numerical methods into ML workflows. His research profile mixes theory and hands-on engineering, from single-cell compartmental models to scalable differentiable solvers, revealing a rare comfort moving between ionic-scale biology and GPU-accelerated ML code.
8 years of coding experience
7 years of employment as a software developer
Bachelor's Degree, Neuroscience and Mathematics, 3.73, Bachelor's Degree, Neuroscience and Mathematics, 3.73 at Unviersity College UtrechtUniversity of California, San DiegoMaster of Science - MS, Neural Information Processing, Master of Science - MS, Neural Information Processing at University of TuebingenDoctor of Philosophy - PhD, Computational Neuroscience and Machine Learning, Doctor of Philosophy - PhD, Computational Neuroscience and Machine Learning at Stony Brook University
English, Polish, French
Github Skills (11)
solver10
differential-equations10
ode-solver10
ode10
neural10
scientific-machine-learning10
odeint10
julia10
callback9
automatic-differentiation9
neural-network8
Programming languages (10)
Github contributions (5)
SciML/DiffEqFlux.jl
Oct 2019 - Nov 2019
Pre-built implicit layer architectures with O(1) backprop, GPUs, and stiff+non-stiff DE solvers, demonstrating scientific machine learning (SciML) and physics-informed machine learning methods
Role in this project:
ML Engineer
ML EngineerContributions:9 commits, 1 PR, 3 comments in 29 days
Contributions summary:Piotr primarily contributed to the `diffeqflux.jl` project by modifying and enhancing the core functionality of neural ordinary differential equations (ODEs). Their work focused on refining the adjoint method for gradient computation and improving the handling of callbacks within the ODE solver framework. They also added and updated documentation for clarity. Furthermore, the user introduced changes related to named tuples and kwargs, likely to improve code robustness and integration with the broader SciML ecosystem.
scientific-machine-learningpartial-differential-equationsdifferentialodescientific-ml
PiotrSokol/DiffEqRNN
Jan 2021 - May 2021
Contributions:36 pushes in 4 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.