Gabriel De Marmiesse
Software Engineer at Kyutai
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolGabriel De Marmiesse is a seasoned software engineer based in Paris with 10 years of experience building and maintaining production-grade ML and developer tooling. He has deep open-source roots as a current maintainer of Python-on-Whales and former maintainer/contributor to high-profile projects like Keras and TensorFlow Addons, where he implemented performance-minded solutions such as lazy loading for shared ops. His background blends backend engineering, DevOps, and ML engineering—improving CI/CD, dependency management, and hyperparameter tuning tooling across projects like Sacred, Keras Tuner, and AutoKeras. Educated in data science and machine learning at Telecom ParisTech and EURECOM, he pairs rigorous academic training with pragmatic fixes and refactors that improve reliability and developer experience. Currently exploring Mojo, he brings a knack for subtle but impactful infrastructure and developer-experience improvements that often go unnoticed until they stop working.
10 years of coding experience
Master's degree, Data science & machine learning, Master's degree, Data science & machine learning at EURECOMMaster's degree, Data Science, Master's degree, Data Science at Telecom ParisTech
Japanese, Spanish, French, English
Stackoverflow
41reputation
4kreached
1answer
0questions
Github Skills (34)
rep10
python10
travis10
testing10
machine-learning10
cicd10
hyperparameter-optimization10
keras10
deep-learning10
tensorflow10
automl10
repr10
cython10
build-automation10
neural-network10
Programming languages (9)
Github contributions (5)
keras-team/keras-contrib
Dec 2018 - Dec 2019
Keras community contributions
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:92 commits, 179 PRs, 93 pushes in 1 year
Contributions summary:Gabriel made contributions to improve the project's build and dependency management. Their work included syncing the Travis configuration with the Keras project, and fixing the installation of dependencies within the build environment. These changes improved the stability and reliability of the continuous integration pipeline.
pythondata-sciencedeep-learningtheanoneural-networks
tensorflow/addons
Jan 2020 - Dec 2021
Role in this project:Back-end Developer
Back-end DeveloperContributions:16 reviews, 310 commits, 683 PRs in 1 year 11 months
Contributions summary:Gabriel implemented lazy loading for shared libraries, improving the efficiency and potentially the startup time of the TensorFlow Addons custom ops. Their work involved modifying core image processing, activation, and other custom ops to utilize this lazy loading mechanism. These changes were applied across multiple files in several submodules of the project.
python2-xdeep-learningaddonssig
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.