Nicolas Hug
ML Research Engineer at Meta
Stony Stratford, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolNicolas Hug is an ML Research Engineer with 12 years of experience building and maintaining core open-source machine learning libraries, currently on the PyTorch team at Meta where he leads development for TorchVision and TorchCodec. He combines deep research training (PhD and postdoc work) with production-grade engineering, focusing on preprocessing and performance optimizations for image and video pipelines. His contributions include CPU-optimized interpolation, uint8 support and channels_last speedups in pytorch/pytorch, AVIF/HEIC and video decoding work in TorchVision, and broad improvements across PyTorch repos from CI to docs and tests. A long-time core contributor to scikit-learn and author of the Surprise recommendation library, he bridges statistical ML, numerical methods (SciPy), and practical tooling for practitioners. Based in the UK, he’s equally comfortable refactoring low-level kernels and polishing user-facing documentation, a balance that drives both reliability and adoption. An understated strength is his habit of improving test coverage and developer experience, making complex libraries easier to use and extend.
12 years of coding experience
2 years of employment as a software developer
Master’s Degree Artificial Intelligence Machine Learning, Master’s Degree Artificial Intelligence Machine Learning at Université Paul Sabatier Toulouse IIIMaster’s Degree Computer Science Engineering, Master’s Degree Computer Science Engineering at INSA Toulouse - Institut National des Sciences Appliquées de Toulouse
French, English, Spanish
Github Skills (68)
recommendation-system10
restructuredtext10
python10
operation10
image-processing10
testing10
refactor10
deep-learning10
ci-cd-pipeline10
refactoring10
scientific-computing10
optimization10
recommendation-engine10
scikit10
statistics10
Programming languages (14)
Github contributions (5)
pytorch/hub
Aug 2021 - Jan 2023
Submission to https://pytorch.org/hub/
Role in this project:
DevOps Engineer
DevOps EngineerContributions:50 reviews, 61 commits, 62 PRs in 1 year 5 months
Contributions summary:Nicolas primarily focused on improving the CI/CD pipeline and build processes for the PyTorch Hub repository. Their contributions include setting up nightly builds using the PyTorch nightly channel, switching from bash to pytest for testing, and fixing CI failures. Additionally, the user modified the installation scripts to include dependencies and to improve their readability, and also brought back a test CI job.
pytorch
scikit-learn/scikit-learn
Nov 2017 - Sep 2021
scikit-learn: machine learning in Python
Role in this project:Data Scientist
Data ScientistContributions:565 reviews, 288 commits, 588 PRs in 3 years 10 months
Contributions summary:Nicolas's contributions primarily involve improving the documentation and the codebase of machine learning algorithms related to the scikit-learn library. They fixed several documentation issues, including typos, and updated mathematical formulations. Furthermore, they added examples related to Principal Component Regression (PCR) and Partial Least Squares Regression (PLS) for better explaining model usages. Their work also focused on improving the code by removing deprecated functions, methods and adjusting their behaviour and parameter values.
data-analysispythonstatisticsdata-sciencelearn-machine-learning
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.