Luca Wehrstedt
Research Engineer (Artificial Intelligence) at Facebook
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolLuca Wehrstedt is a research engineer specializing in applied AI, with 14 years of software engineering experience and a strong theoretical computer science background from École normale supérieure. Based in Paris, he builds and optimizes state-of-the-art ML systems at Meta (Facebook), contributing notable performance and low-level CUDA/PyTorch improvements across projects like xformers and PyTorch itself. His work blends backend systems engineering, numerical performance tuning (including FP8 and FlashAttention), and maintainability-driven refactors for large-scale graph embedding and contest management tools. Luca’s profile combines industrial research, top-tier academic training, open-source impact, and a taste for programming competitions—an indicator of both deep technical rigor and practical problem-solving.
14 years of coding experience
1 year of employment as a software developer
Bachelor's degree, Bachelor's degree at Ecole normale supérieureUniversity of BolognaExchange student at the Graduate School of Arts and Sciences, Exchange student at the Graduate School of Arts and Sciences at Columbia University in the City of New YorkDiploma, Diploma at Collegio Superiore di Bologna
Italian, German, English, French
Github Skills (25)
pytorch10
debug10
c-language10
postgresql10
python10
operation10
api-design10
tensorrt10
machine-learning10
tensorflow10
sqlalchemy10
gpu10
performance-optimization10
cuda10
tensor10
Programming languages (9)
Github contributions (5)
facebookresearch/PyTorch-BigGraph
Apr 2019 - Oct 2021
Generate embeddings from large-scale graph-structured data.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:2 releases, 9 reviews, 77 commits in 2 years 7 months
Contributions summary:Luca primarily contributed to bug fixes and refactoring within the PyTorch-BigGraph repository. Their work involved resolving issues related to parameter handling and data export functionalities, particularly within dynamic relations. Furthermore, the user updated documentation to clarify information previously discussed in GitHub issues and addressed compatibility problems with Python 3.7. These changes show a focus on code correctness and maintainability within the project.
structuredgnnembeddingsgraph-neural-networkmachine-learning
cms-dev/cms
Sep 2011 - Mar 2019
Contest Management System
Role in this project:Back-end Developer
Back-end DeveloperContributions:1 release, 843 commits, 194 PRs in 7 years 6 months
Contributions summary:Luca's contributions focus on enhancing the Contest Management System's back-end functionalities. Their work includes refactoring core components such as the TaskType class and the user authentication system, introducing features like enhanced token management, and improving the database schema with features such as custom PostgreSQL domains. The user also addressed minor cosmetic and documentation improvements throughout the project.
management-systemprogramming-competitionspythongrading-systemcontest-management-system
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.