Olivier Grisel is a senior machine learning engineer with 17 years of experience, a long-time core contributor to scikit-learn and a founding engineer at :probabl. who bridges research-grade algorithms and production-ready infrastructure. He balances hands-on ML model work—ranging from convolutional networks for image tasks to scalable parallel pipelines—with deep contributions to Python ecosystem tooling such as joblib, cloudpickle, manylinux/auditwheel and OpenBLAS. A frequent conference speaker and MOOC instructor, he also organizes PyData Paris and crafts tutorials that make complex predictive modeling accessible. His open-source impact spans bug fixes to performance and CI improvements, showing a rare mix of algorithmic insight and systems-level engineering. Based in Vannes, France, he combines academic training from ENSTA and Imperial College with practical experience tuning libraries used across the scientific Python stack. Unexpectedly, his contributions include low-level resilience fixes (pthread_atfork) and CI automation that materially improve reliability for large-scale Python deployments.
17 years of coding experience
19 years of employment as a software developer
MSc. in Advanced Computing, Computer Science, MSc. in Advanced Computing, Computer Science at Imperial College London
Ingénieur, Computer Science, Ingénieur, Computer Science at ENSTA
Classes Préparatoires aux Grandes Ecoles, Mathematics and Physics, Classes Préparatoires aux Grandes Ecoles, Mathematics and Physics at Lycée Ste Geneviève
Contributions summary:Olivier's contributions involve implementing and experimenting with unsupervised feature extraction techniques, specifically focused on Gaussian mixture models (GMM) and various neural network architectures, particularly an MLP and possibly an autoencoder variant, for the purpose of performing classification tasks on the MNIST dataset. The user is also working on exploring and improving these methods by testing the performance of various optimizers and examining the effects of learning rates on model performance, analyzing the behavior of different configurations. The primary focus is on understanding how well these models generalize to unseen data.
Contributions:3 releases, 93 reviews, 147 commits in 7 years 10 months
Contributions summary:Olivier primarily contributed to the `cloudpickle` library, enhancing its functionality and maintainability. They added tests for nested constructs, ensuring comprehensive coverage for various pickling scenarios. The user also performed code style improvements and addressed compatibility issues with PyPy3. Furthermore, they addressed issues related to interactive function pickling, and implemented tests for subprocess-based and memoryview-related tests.
pythonpythonicextended
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.