Mohamed Mekkouri
ML Research Engineer at White Circle
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolMohamed Mekkouri is an ML research engineer with four years of hands-on experience at the intersection of machine learning, security, and blockchain, currently based in Paris. Trained at CentraleSupélec and Université de Rennes I, he blends rigorous academic foundations with applied research from internships at CEA and projects with Wavestone on CAPTCHA-breaking computer vision. At Hugging Face he contributed to the widely used transformers library, implementing and testing advanced quantization methods (BitNet, torchao, GGML, AQLM, VQTP, FP8) and shipping quantization kernels that improve model efficiency. He has practical blockchain and Web3 experience—writing and deploying ERC-1155 smart contracts for NFT/FNFT patents—and a knack for turning research prototypes into production-ready components. Now at White Circle, he continues to push efficient ML tooling while keeping a strong interest in cybersecurity and decentralized systems. Colleagues value him for combining low-level ML engineering with a security-minded perspective that surfaces non-obvious vulnerabilities in deployed systems.
4 years of coding experience
3 years of employment as a software developer
BSc. and MSc in Engineering, BSc. and MSc in Engineering at CentraleSupélecMaster of Computer Science, Master of Computer Science at Université de Rennes IPreparatory classes for french engineering schools, MPSI/MP*, Mathematics, Physics and Engineering, Preparatory classes for french engineering schools, MPSI/MP*, Mathematics, Physics and Engineering at Lycée Mohammed VI d'Excellence
Arabic, French, English, German
Github Skills (14)
quantization10
pytorch10
quants10
machine-learning10
deeplearning-ai10
transformer10
nlp10
deep-learning10
python10
triton9
gff8
gfs8
gf8
p87
Programming languages (7)
Github contributions (5)
huggingface/transformers
Aug 2024 - Apr 2025
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Role in this project:
ML Engineer
ML EngineerContributions:79 reviews, 86 PRs, 193 pushes in 8 months
Contributions summary:Mohamed's primary contributions focused on integrating and testing quantization methods within the Hugging Face Transformers library, specifically for BitNet, and other quantization techniques like torchao, GGML, AQLM, VQTP, and FP8. They implemented quantization kernels, debugged and fixed tests related to these methods, and made adjustments to support these quantization techniques in the library's architecture. The user also worked on integrating, testing, and deprecating different quantization methods.
pythonbertspeech-recognitionstate-of-the-artflax
MekkCyber/transformers
Sep 2024 - Mar 2025
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Contributions:61 pushes, 12 branches in 6 months
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.