Patrick Labatut
Research Engineering Manager at Meta
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolPatrick Labatut is a research engineering manager based in Paris who blends deep academic training (PhD in computer science) with hands-on leadership in computer vision, graphics and ML across product-scale systems. He has driven R&D-to-product transitions at Meta (FAIR and Meta Spark) and led large mapping and reality-capture pipelines at HERE, delivering production-grade processing for million-kilometer datasets and hundreds of millions of AR users. Patrick pairs technical depth in rendering, sensor calibration and model efficiency with strong QA and automation instincts—evidenced by contributions to major open-source projects like PyTorch3D, Detectron2 and DINOv2 where he focused on rendering fixes, benchmarks, dataset hygiene and robust testing. He excels at taking early-stage prototypes through maturation, emphasizing performance, reproducibility and maintainability. Colleagues rely on him for pragmatic engineering leadership that bridges research novelty and deployable systems.
6 years of coding experience
14 years of employment as a software developer
B.Sc., M.Sc., Computer Science, Mathematics, B.Sc., M.Sc., Computer Science, Mathematics at Ecole normale supérieureM.Sc., Applied Mathematics, Computer Vision, Machine Learning, M.Sc., Applied Mathematics, Computer Vision, Machine Learning at École normale supérieure Paris-SaclayPh.D., Computer Science, Ph.D., Computer Science at Université Denis Diderot (Paris VII)
English, French, German
Github Skills (16)
testing10
unit-testing10
computer-vision10
pytorch10
machine-learning10
detectron10
software-quality10
python10
datasets10
test-automation10
imagenet9
mesh-processing9
configuration-management9
deep-learning8
depth-estimation8
Programming languages (2)
Github contributions (5)
facebookresearch/dinov2
Apr 2023 - Feb 2024
PyTorch code and models for the DINOv2 self-supervised learning method.
Role in this project:
ML Engineer
ML EngineerContributions:14 reviews, 50 PRs, 48 pushes in 10 months
Contributions summary:Patrick primarily contributed to improving and maintaining the DINOv2 codebase, focusing on dataset preparation and integration. Their work included cleaning up dataset interfaces, fixing issues in the ImageNet-1k dataset, and adding code for depth estimation and semantic segmentation. They also addressed bugs in the linear classifier wrapper and incorporated features related to xFormers integration.
facebookresearch/pytorch3d
Feb 2020 - Nov 2021
PyTorch3D is FAIR's library of reusable components for deep learning with 3D data
Role in this project:Backend Developer & Test Automation Engineer
Backend Developer & Test Automation EngineerContributions:2 reviews, 46 commits, 1 push in 1 year 9 months
Contributions summary:Patrick primarily contributed to the codebase by fixing spelling errors, indicating a focus on code quality and detail. They also made changes to the code related to various shaders (Gouraud, Phong, and Silhouette) demonstrating an understanding of the rendering pipeline. Additionally, the user was involved in modifying and adding benchmarks for mesh I/O operations, which suggests a role in performance testing and optimization of the project.
pytorchdeep-learningcomputer-graphicscomputer-vision3d-data
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.