Julien Cuquemelle
SRE At GPU-Platform
Paris, Ile-de-France
Summary
👤
Senior🎓
Top SchoolJulien Cuquemelle is an experienced SRE and ML infrastructure engineer with 11+ years building high-performance GPU and multicore systems for production AI at Criteo and earlier R&D roles. He specializes in industrializing GPU-based training and Triton-powered inference, driving migrations from Mesos/YARN to Kubernetes and delivering major cost and CO2 reductions by moving workloads to GPU. Julien combines deep low-level experience (C++, C#, multicore/GPU) with big-data expertise (Spark, Hadoop) and has contributed backend improvements to the popular MLflow project, optimizing SQL-backed filtering and pagination. He’s also a practical knowledge-sharer who authors tooling, benchmarks and best practices that shaped hardware refresh and platform roadmaps, and who advocates ONNX to decouple training and inference. Based in Paris, his background in robotics/AI and sensory analysis gives him an uncommon cross-disciplinary view of applied ML systems.
11 years of coding experience
18 years of employment as a software developer
Automotive technology, Automotive technology at Technische Universität BerlinMSc Robotics AI, MSc Robotics AI at Université de Technologie de Compiègne (UTC)
French, English, German
Github Skills (14)
mlflow10
sqlalchemy10
sql10
python10
pagination9
testing9
unit-test9
machine-learning9
unit-testing9
apache-spark8
git8
docker6
dockers6
cicd6
Programming languages (4)
Github contributions (5)
mlflow/mlflow
Oct 2019 - Mar 2020
Open source platform for the machine learning lifecycle
Role in this project:
Back-end Engineer
Back-end EngineerContributions:61 commits, 8 PRs, 9 comments in 5 months
Contributions summary:Julien contributed significantly to the back-end functionality of the MLflow project. They focused on pushing filter functions from Python to SQL, optimizing SQL queries, and generalizing ordering and pagination. Their work involved modifying tests related to large data and run searching, while also fixing issues related to Protobuf generation. The user also fixed a bug related to pip versions.
pythonlifecyclemlmachine-learningincremental-learning
criteo/cluster-pack
Oct 2020 - Jan 2023
A library on top of either pex or conda-pack to make your Python code easily available on a cluster
Contributions:30 reviews, 43 commits, 30 PRs in 2 years 3 months
python-codepexpythoncondaconda-pack
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.