Duncan Zauss
Senior Software Engineer at Meta
Zurich, Zurich, Switzerland
Summary
👤
Senior🎓
Top SchoolDuncan Zauss is a Senior Software Engineer based in Zurich with six years of experience building computer vision, deep learning, and robotics systems—currently scaling data and training pipelines for 3D generative AI at Meta. He has driven NeRF- and Gaussian-splatting-based photorealistic metaverse efforts and previously worked on 3D scene reconstruction at Meta Reality Labs, bridging research ideas to production. His academic background includes a strong CV focus with a first-author ICCV paper from work at EPFL and an honors M.Sc. in Computational Engineering from TU Darmstadt. An active contributor to open-source ML, he implemented weighted loss functions and whole-body keypoint support in the well-known OpenPifPaf project, showing both algorithmic and engineering chops. Duncan combines real-time inference experience (YOLO/Faster-RCNN) and research-grade NeRF expertise, making him effective at shipping scalable vision models. Colleagues value his blend of rigorous academic training and pragmatic production engineering across large-scale ML systems.
6 years of coding experience
4 years of employment as a software developer
Abitur (German A-Levels), 1.4, Award for outstanding achievements in Physics, Abitur (German A-Levels), 1.4, Award for outstanding achievements in Physics at Otto-Hahn-Gymnasium SpringeTechnischen Universität DarmstadtTechnical University of DenmarkBachelor of Engineering, Mechanical Engineering, 1.9, Bachelor of Engineering, Mechanical Engineering, 1.9 at Duale Hochschule Baden-WürttembergSemester abroad - Master thesis at the Visual Intelligence for Transportation (VITA) Lab, Computer Vision, 1.0 (German grade scale), Paper accepted at ICCV '21 (First author), Semester abroad - Master thesis at the Visual Intelligence for Transportation (VITA) Lab, Computer Vision, 1.0 (German grade scale), Paper accepted at ICCV '21 (First author) at EPFL (École polytechnique fédérale de Lausanne)Universitat Politècnica de València
German, English, French, Spanish
Github Skills (6)
computer-vision10
pytorch10
deep-learning10
human-pose-estimation10
python9
machine-learning9
Programming languages (1)
Github contributions (5)
openpifpaf/openpifpaf
Mar 2021 - Aug 2021
Official implementation of "OpenPifPaf: Composite Fields for Semantic Keypoint Detection and Spatio-Temporal Association" in PyTorch.
Role in this project:
ML Engineer
ML EngineerContributions:10 reviews, 13 commits, 13 PRs in 5 months
Contributions summary:Duncan implemented a weighted loss function within the OpenPifPaf framework. This involved modifying existing loss functions and integrating the training weights into the core loss calculation. The user also added support for the Apollo 66 keypoints within the apollocar3d plugin and included a new checkpoint. Finally, the user added the wholebody and weighting features based on local centrality metrics.
pytorchkeypoint-detectionopenpifpafassociationdeep-learning
vita-epfl/openpifpaf_wholebody
Dec 2020 - Jul 2021
Contributions:30 commits, 3 pushes 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.