Kevin Chen
Staff Machine Learning Engineer at Matterport
San Jose, California, United States
Summary
🤩
Rockstar🎓
Top SchoolKevin Chen is a Staff Machine Learning Engineer in San Jose with five years of industry experience applying deep learning to 3D computer vision and NeRF-based systems. At Matterport he progressed from computer vision engineer to staff-level ML lead, shipping production features and model improvements that bridge research and scalable product needs. His open-source contributions to the widely used nerfstudio project include fixes and enhancements to NeRF training, camera pose handling, and core MLP/spherical harmonics code, reflecting comfort with both model internals and rendering pipelines. Kevin holds a PhD in Physics from UCLA and a background in applied math and physics from UC Berkeley, a combination that gives him a strong theoretical footing for tackling practical ML engineering problems. Notably, his trajectory spans fundamental research in theoretical physics to hands-on engineering at scale, enabling him to translate complex concepts into reliable systems.
5 years of coding experience
9 years of employment as a software developer
Bachelor’s Degree Physics; Applied Mathematics, Bachelor’s Degree Physics; Applied Mathematics at University of California, BerkeleyUniversity of California, Los Angeles
Github Skills (9)
pytorch10
machine-learning10
computer-vision10
serf10
python10
3d-reconstruction10
3d9
mlp9
3d-graphics8
Programming languages (7)
Github contributions (5)
nerfstudio-project/nerfstudio
Oct 2022 - Nov 2022
A collaboration friendly studio for NeRFs
Role in this project:
ML Engineer
ML EngineerContributions:19 reviews, 16 commits, 17 PRs in 1 month
Contributions summary:Kevin primarily contributed to the `nerfstudio` project by fixing bugs and adding new features related to NeRF model training and viewer functionality. They addressed issues in camera pose handling, depth normalization, and image processing, specifically for image resampling and limiting images in the viewer to improve performance. Furthermore, the user made modifications to the MLP and spherical harmonics calculations, suggesting involvement with the core neural network components of the project.
pytorchphotogrammetrydeep-learningcomputer-vision3d-reconstruction
kevinddchen/Keras-FCN
Mar 2021 - Dec 2021
Keras implementation of fully convolutional network for semantic image segmentation on the Pascal VOC dataset. Based on Shelhamer et al. (2016).
Contributions:21 commits, 5 PRs, 41 pushes in 9 months
pascalpascal-vocvoc-datasetsemantic-segmentationdeep-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.