Xavier Dupré
Computer Scientist at Microsoft
Paris, Ile-de-France
Summary
🤩
RockstarXavier Dupré is a computer scientist based in Paris with a PhD and over a decade of hands-on experience bridging ML research and production tooling. He has a strong track record contributing to flagship open-source projects in the ONNX ecosystem and PyTorch, improving model conversion, exporter robustness, and Python bindings to make ML models more interoperable and deployable. His work spans operator-level fixes (FFT, WordpieceTokenizer), nested subgraph handling for torch.cond, and enhancing runtimes like onnxruntime—skills that combine deep framework knowledge with practical engineering. A former Microsoft engineer and longtime teacher at ENSAE, he pairs industry experience with academic rigor and maintains an active teaching account (sdpython). Colleagues value him for quietly solving subtle interoperability and CI problems that enable others to ship models reliably.
11 years of coding experience
Github Skills (36)
pytorch10
c-language10
python10
apidoc10
scikit10
machine-learning10
onnx10
numpy10
keras10
deep-learning10
api10
tensorflow10
computer-engineering10
ci-cd-pipeline10
scikit-learn10
Programming languages (7)
Github contributions (5)
onnx/onnxmltools
Mar 2018 - Jan 2023
ONNXMLTools enables conversion of models to ONNX
Role in this project:
DevOps Engineer & ML Engineer
DevOps Engineer & ML EngineerContributions:5 releases, 55 reviews, 58 commits in 4 years 11 months
Contributions summary:Xavier primarily focused on improving the continuous integration (CI) setup and converting machine learning models to ONNX format. Their commits include setting up and configuring CI/CD pipelines using tools like Appveyor, Travis CI, and CircleCI, indicating a focus on automating builds and tests. They also addressed issues related to model conversion, including fixing dependencies, adding CoreML tools, and addressing specific model conversion bugs, demonstrating ML engineering involvement. Furthermore, the user worked to enable support for running tests with onnxruntime.
onnxmachine-learningpython-libraryscikit-learnkeras
onnx/sklearn-onnx
Jan 2019 - Jan 2023
Convert scikit-learn models and pipelines to ONNX
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:22 releases, 107 reviews, 365 commits in 4 years 1 month
Contributions summary:Xavier contributed to the refactoring of the codebase, renaming files and classes to improve readability. They were involved in converting scikit-learn models to ONNX format. The commits also include examples demonstrating the use of custom operators for Gaussian Mixture models. They made improvements to the accuracy and functionalities of the model.
onnxscikit-learn
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.