Jérémy Rapin
Research Engineer at Meta
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolJérémy Rapin is a research engineer at Meta AI with 11 years of experience building data and training infrastructure for large-scale ML research. He is a founding maintainer of Nevergrad and the original creator of submitit, contributing deep expertise in optimization, job scheduling on Slurm, and robust backend tooling. His background spans applied mathematics and signal processing (PhD), and he led production-grade deep learning efforts in medical AI that achieved regulatory milestones like CE/FDA pathways. Comfortable bridging research and production, he has shipped tested Python libraries, designed structured configuration and optimization plugins for Hydra, and improved cluster automation. Based in Paris, he pairs rigorous academic training with hands-on systems engineering, often working on the low-level algorithmic details that make large ML experiments reproducible and scalable. A less obvious strength is his track record of turning niche research code into widely used open-source infrastructure relied on across Meta and the broader ML community.
11 years of coding experience
Engineer, Applied Mathematics (image processing, statistics), Engineer, Applied Mathematics (image processing, statistics) at Ecole centrale de ParisMaster 2 (M2), Applied Mathematics, Master 2 (M2), Applied Mathematics at École normale supérieure Paris-SaclayDoctor of Philosophy (PhD), Mathématiques appliquées, Doctor of Philosophy (PhD), Mathématiques appliquées at Université Paris Sud (Paris XI)Master of Science, Applied Mathematics (image processing, statistics), Master of Science, Applied Mathematics (image processing, statistics) at Ecole normale supérieure de Cachan
French, English, Spanish, langue des signes française
Github Skills (16)
algorithm10
configuration-management10
cluster-computing10
slurm10
python10
optimizers10
optimisation10
clustering10
optimization10
testing10
devops9
automation9
automations9
unit-testing9
machine-learning8
Programming languages (6)
Github contributions (5)
facebookincubator/submitit
May 2020 - Sep 2022
Python 3.8+ toolbox for submitting jobs to Slurm
Role in this project:
Back-end & DevOps Engineer
Back-end & DevOps EngineerContributions:73 reviews, 42 commits, 92 PRs in 2 years 4 months
Contributions summary:Jérémy primarily contributed to the automation and improvement of job submission processes for a Slurm cluster. They focused on enhancing the `submitit` toolbox, a Python library for submitting jobs. Their commits include adding environment variables for executor detection, modifying the code for cluster-specific overrides, and improving the internal structure of the `AutoExecutor`. Additionally, the user addressed linter issues, updated documentation, and made code formatting changes to enhance readability and maintainability.
slurmpythonorchestrationclusterspython-3-6
facebookresearch/nevergrad
Dec 2018 - May 2022
A Python toolbox for performing gradient-free optimization
Role in this project:Back-end Developer
Back-end DeveloperContributions:16 releases, 886 reviews, 717 commits in 3 years 5 months
Contributions summary:Jérémy primarily focused on fixing errors within the `nevergrad/optimization` module. These commits involved changes to files related to optimization algorithms, including code modifications and corrections, indicating a focus on improving the core functionality and stability of the library. The changes include modifications in `recastlib.py`, `corefuncs.py`, and `test_optimizerlib.py`, which suggest work on optimization models, their performance, and testing infrastructure.
pythontoolboxgradient-free-optimizationoptimizationpython-toolbox
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.