Dario Pavllo
Machine Learning Engineer at QuantCo
Zurich, Zurich, Switzerland
Summary
👤
Senior🎓
Top SchoolDario Pavllo is a machine learning engineer and recent ETH Zürich PhD graduate with eight years of experience bridging deep learning research and production engineering. He has interned at Google and Facebook AI Research working on NeRFs, single-view 3D reconstruction, and video-based human motion modeling, and contributed practical improvements to the widely used VideoPose3D project (Detectron2 integration, inference-in-the-wild support, and visualization fixes). At ETH he led large-scale teaching for deep learning courses and supervised student projects tied to funded research, and now applies that expertise at QuantCo. His background spans software engineering, distributed systems, and security, with early industry experience building C++/GPU game systems and procedural content tools shipped on Steam. Colleagues describe him as a researcher who prioritizes reproducible, usable ML systems—turning academic models into robust pipelines for real-world inference.
8 years of coding experience
10 years of employment as a software developer
University of BolognaMaster of Science - MS, Computer Science, Master of Science - MS, Computer Science at Ecole polytechnique fédérale de LausanneIstituto Tecnico Industriale "Blaise Pascal", CesenaDoctor of Philosophy - PhD, Machine Learning, Doctor of Philosophy - PhD, Machine Learning at ETH Zurich
German, French, English, Italian
Github Skills (11)
computer-vision10
pytorch10
machine-learning10
inference10
pose-detection10
python10
pose-estimation10
3d10
numpy9
detectron9
matplotlib8
Programming languages (2)
Github contributions (5)
facebookresearch/VideoPose3D
Nov 2018 - Aug 2020
Efficient 3D human pose estimation in video using 2D keypoint trajectories
Role in this project:
ML Engineer
ML EngineerContributions:10 commits, 3 PRs, 113 comments in 1 year 9 months
Contributions summary:Dario's contributions focused on enhancing the video pose estimation pipeline. They addressed data loading issues related to NumPy versions and added support for inference in the wild, including integration with Detectron2. Further improvements involved incorporating trajectory support during inference, enhancing the model's capabilities and usability. The user also fixed the visualization script for compatibility with newer Matplotlib versions, ensuring accurate results.
human-poseunsupervised-learningcomputer-vision3d-human-posepose-estimation
epfl-dlab/quootstrap
Dec 2017 - Jul 2018
Contributions:1 release, 26 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.