Ilyas Moutawwakil
Machine Learning Optimization Engineer
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolIlyas Moutawwakil is a Machine Learning Optimization Engineer based in Paris with six years of experience specializing in ML framework-hardware co-optimization, quantization, and performance benchmarking. At Hugging Face he drives multi-vendor hardware integrations (AMD, Intel, Habana) and maintains Optimum-Benchmark and the LLM Perf Leaderboard to measure latency, throughput, memory and energy across backends. He contributes to flagship open-source projects like Hugging Face Transformers—adding Exllama/ AWQ kernel support, ONNX export fixes, HPU work and cache improvements—bridging research-quality models with production constraints. His background blends academic training in optimization and data science (CentraleSupélec, Université de Lorraine) with applied work on OCR, document understanding and table reconstruction for industry customers. Colleagues rely on him to translate low-level kernel and quantization challenges into reproducible OSS integrations that scale across devices.
6 years of coding experience
2 years of employment as a software developer
Master of Engineering - MEng Data & Information Science , Master of Engineering - MEng Data & Information Science at CentraleSupélecMaster of Science - MS Optimization, Master of Science - MS Optimization at Universite de LorraineAssociate's degree Mathematics Physics and Engineering, Associate's degree Mathematics Physics and Engineering at Association pour la Promotion de l'Enseignement d'Excellence
French, English, Arabic, Japanese
Stackoverflow
89reputation
8kreached
7answers
0questions
Github Skills (19)
transformers10
pytorch10
python10
machine-learning10
deep-learning10
nlp10
quantization9
onnx8
camera6
data-analysis6
scikit-learn6
plotly6
geopandas6
opencv6
dataframe6
Programming languages (8)
Github contributions (5)
huggingface/transformers
Apr 2022 - Apr 2025
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Role in this project:
ML Engineer
ML EngineerContributions:79 reviews, 15 PRs, 81 pushes in 3 years
Contributions summary:Ilyas primarily contributed to the integration of the Exllama kernels for AWQ models within the Hugging Face Transformers library. Their work involved adding support for new AWQLinear versions and Exllama configurations, including version checks and integration within the quantization process. They also addressed issues related to ONNX exportability, specifically fixing non-exportable inplace operations. Furthermore, the user made contributions related to HPU support, and also made cache improvements.
pythonbertspeech-recognitionstate-of-the-artflax
IlyasMoutawwakil/cyclegan
Sep 2020 - Jun 2022
Contributions:8 commits, 7 pushes, 1 branch in 1 year 8 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.