Leandro Von Werra
Head Of Research at Hugging Face
Bern, Bern, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolLeandro Von Werra is Head of Research at Hugging Face with a decade of experience bridging NLP research and production-grade machine learning. He has driven RL-for-language work—implementing PPO for transformers—and added widely used evaluation tooling and datasets (MBPP, HumanEval, pass@k) to Hugging Face’s core libraries. Comfortable shipping both research code and robust engineering (stream processing, dataset integrations, custom generation hooks), he taught data science at BFH and previously applied NLP and time-series solutions in industry. Trained as a physicist at ETH Zürich, he pairs rigorous quantitative thinking with hands-on open-source contributions that materially improve model evaluation and safety in real-world ML stacks.
10 years of coding experience
8 years of employment as a software developer
Master of Science - MS Physics, Master of Science - MS Physics at ETH Zürich
Github Skills (38)
transformers10
code-optimization10
pytorch10
ppp10
language-model10
python10
evaluation10
kafka10
machine-learning10
datasets10
asyncio10
transformer-models10
deeplearning-ai10
tokenizer10
code-generation10
Programming languages (8)
Github contributions (5)
nlp-with-transformers/notebooks
Jan 2022 - Jan 2023
Jupyter notebooks for the Natural Language Processing with Transformers book
Role in this project:
ML Engineer
ML EngineerContributions:11 reviews, 20 commits, 8 PRs in 11 months
Contributions summary:Leandro primarily contributed to a Jupyter notebook repository focused on Natural Language Processing with Transformers. Their work involved fixing import statements and adding notes related to model loading and Colab/Kaggle compatibility. They also added and removed code related to visualizations and large datasets. The user further added code to install the `wandb` library, demonstrating a focus on model training and evaluation within this domain. Finally, they included and updated figures in the chapters to improve the user's experience.
nlptransformerspythonlanguage-processingnatural-language
huggingface/evaluate
Nov 2021 - Dec 2022
🤗 Evaluate: A library for easily evaluating machine learning models and datasets.
Role in this project:ML Engineer
ML EngineerContributions:7 releases, 323 reviews, 189 commits in 1 year 1 month
Contributions summary:Leandro's primary contribution revolves around implementing and integrating the OpenAI's pass@k code evaluation metric. Their work includes adding the metric code, optimizing code style, handling potential multiprocessing issues, and improving the metric's documentation. They also addressed platform-specific issues, such as excluding the metric from Windows tests, and incorporated copyright and licensing information. The user's modifications involve changes to the core `code_eval` metric implementation.
evaluateevaluation-frameworkdata-scienceevaluationmachine-learning
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.