Gabriel Synnaeve
Research Scientist at Facebook
New York, New York, France
Summary
🤩
Rockstar🎓
Top SchoolGabriel Synnaeve is a research scientist with 17 years of experience blending Bayesian modeling, deep learning, and practical ML engineering across industry and academia. Based in New York and working at Facebook, he brings a rare combination of PhD‑level research in Bayesian approaches to games and hands‑on engineering contributions to prominent open‑source projects like scikit‑learn (RBM implementation and testing) and the C++ flashlight library (transformer position embeddings, training schedules, SpecAugment). His background spans speech acquisition, end‑to‑end speech recognition, weak supervision, and real‑time strategy game AI, reflecting both theoretical depth and applied system design. Notably, he has a track record of modernizing legacy C++ code (memory fixes and refactors) and shipping reproducible ML components that bridge research prototypes and production tooling.
17 years of coding experience
Mathematics, Physics, Engineering, Mathematics, Physics, Engineering at Lycée Saint LouisDoctor of Philosophy (PhD), Bayesian modeling, machine learning, Doctor of Philosophy (PhD), Bayesian modeling, machine learning at Grenoble UniversityMaster, Mathematics and Computer Science, Master, Mathematics and Computer Science at Ecole Nationale Supérieure d'Informatique et de Mathématiques Appliquées de Grenoble / ENSIMAGMaster, Artificial Intelligence and Web, Master, Artificial Intelligence and Web at Université Joseph Fourier (Grenoble I)
French, English, German
Stackoverflow
646reputation
2kreached
2answers
0questions
Github Skills (24)
c-language10
python10
scikit10
machine-learning10
numpy10
transformer-models10
deep-learning10
scikit-learn10
neural-network10
autograd10
cprogramming-language10
data-science9
testing9
pytorch9
data-analysis8
Programming languages (9)
Github contributions (5)
TorchCraft/TorchCraft
Dec 2016 - Nov 2018
Connecting Torch to StarCraft
Role in this project:
Back-end Developer
Back-end DeveloperContributions:33 commits, 19 PRs, 44 pushes in 1 year 11 months
Contributions summary:Gabriel primarily contributed to the back-end logic of the project, focusing on improvements to existing StarCraft integration using the BWAPI library. Their work involved fixing memory leaks, refactoring code to use modern C++ constructs like `delete[]`, and adding features. They also added helper functions for unit production costs. Furthermore, the user addressed issues related to upgrades, technologies and races, ensuring correct data representation and alignment.
starcraftminecraftbwapideep-learningreinforcement-learning
scikit-learn/scikit-learn
Oct 2012 - Jul 2013
scikit-learn: machine learning in Python
Role in this project:ML Engineer
ML EngineerContributions:23 commits in 9 months
Contributions summary:Gabriel primarily contributed to the implementation and testing of a Bernoulli Restricted Boltzmann Machine (RBM) within the scikit-learn library. Their work involved refactoring RBM code, adding tests for Gibbs sampling and sparse matrix inputs, and ensuring correct pseudo-likelihood calculations. Furthermore, they updated an example to showcase the RBM's use in a classification task.
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.