Mickael Ide
Senior Machine Learning Infrastructure Engineer at NVIDIA
Zurich, Zurich, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolMickael Ide is a Senior Machine Learning Infrastructure Engineer based in Zurich with 11 years of experience building high-performance ML systems and CUDA-accelerated libraries. Now at NVIDIA, he focuses on production-grade ML infrastructure and has a strong open-source track record contributing to prominent RAPIDS and CuPy projects—fixing PCA implementations, adding distance metrics, and implementing robust sparse-matrix fallbacks. His background blends low-level C++/CUDA engineering with applied ML (medical imaging segmentation and BI deployments), enabling him to bridge algorithmic research and production performance. Notably, he promoted IncrementalPCA from experimental in cuML and optimized core numerical kernels in RAFT, highlighting a knack for stabilizing and scaling advanced ML primitives.
11 years of coding experience
5 years of employment as a software developer
Master of Science, Data Science & Machine Learning, Master of Science, Data Science & Machine Learning at EPITA: Ingénierie Informatique
French, English
Github Skills (33)
algorithm10
distance-matrix10
algorithms10
sparse-array10
c-language10
cublas10
gpgpu10
numerical10
pytest10
computation10
python10
sparse-matrix10
machine-learning10
machine-learning-algorithms10
sparse-file10
Programming languages (5)
Github contributions (5)
rapidsai/raft
Jan 2021 - Jan 2023
RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing high performance applications.
Role in this project:
Back-end Developer / ML Engineer
Back-end Developer / ML EngineerContributions:118 reviews, 71 commits, 62 PRs in 2 years
Contributions summary:Mickael primarily contributed to the `raft` repository by implementing and correcting distance calculation algorithms. They added new distance metrics like Jensen-Shannon and BrayCurtis to the `DistanceType` enum and corrected the Dice formula. They also addressed NaN issues in Hellinger distance calculations and optimized the code by utilizing 64-bit CuSolver API for Eigen decomposition. Furthermore, the user unified weighted mean code, fixed pointer arithmetic, and improved data structures within the code, making significant contributions to core numerical and mathematical functionalities.
sciencedata-sciencemachine-learninggraph-data-sciencefundamental
rapidsai/cuml
Dec 2020 - Jan 2023
cuML - RAPIDS Machine Learning Library
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:106 reviews, 72 commits, 68 PRs in 2 years 2 months
Contributions summary:Mickael's contributions center around enhancing the `cuML` library, a RAPIDS machine learning library, by addressing and resolving issues related to the Principal Component Analysis (PCA) and incremental PCA algorithms. This includes fixing attribute errors, correcting input types for PCA, and adding tests to validate the fixes. Furthermore, the user promoted `IncrementalPCA` from experimental. Additionally, the user contributed to making the library compatible with different input types and updating documentation.
cudacumlnvidiadata-sciencegpu
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.