Yen Lin
Research Scientist at Google DeepMind
United States
Summary
🤩
Rockstar🎓
Top SchoolYen Lin is a research scientist with 11 years of experience building generative and neural graphics models, currently at Google DeepMind after contributing to NVIDIA’s Cosmos research. She has deep hands-on expertise in NeRF and image-to-image translation, evidenced by contributions to high-profile open-source projects like instant-ngp and NeRF-PyTorch where she improved data preprocessing, camera handling, and training interfaces. Yen’s background spans both research and engineering: she fixed core TensorFlow pix2pix bugs, hardened adversarial-attack utilities in cleverhans, and improved scikit-learn documentation and examples from her earlier Google Summer of Code work. With a PhD in EECS from MIT and a BS in Computer Science, she blends rigorous academic training with pragmatic engineering that focuses on dataset pipelines and reproducible model training. Notably, her contributions often center on the less-visible but critical plumbing—data loaders, transform normalization, and test infrastructure—that makes advanced models reliable at scale.
11 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD EECS, Doctor of Philosophy - PhD EECS at Massachusetts Institute of TechnologyBachelor of Science (BS) Computer Science, Bachelor of Science (BS) Computer Science at National Tsing Hua University
Github Skills (30)
pytorch10
python10
scikit10
machine-learning10
data-preprocessing10
deep-learning10
tensorflow10
scikit-learn10
neural-network10
computer-vision10
modeling10
3d-reconstruction10
data-science9
model-driven9
detectron9
Programming languages (13)
Github contributions (5)
yenchenlin/nerf-pytorch
Mar 2020 - Jul 2022
A PyTorch implementation of NeRF (Neural Radiance Fields) that reproduces the results.
Role in this project:
ML Engineer
ML EngineerContributions:88 commits, 15 PRs, 60 pushes in 2 years 4 months
Contributions summary:Yen made multiple contributions to the NeRF-PyTorch project, focusing on improving the training interface and adding new functionalities. Their commits involved adding documentation strings, implementing center cropping during training, and fixing an intrinsic problem, suggesting a focus on refining the model's training process and data handling. The user also updated the loading of Blender data, indicating involvement in data preprocessing. These modifications and optimizations likely improved the model's performance.
pytorchdeep-learningneural-radiance-fieldscomputer-graphicsvolume-rendering
yenchenlin/pix2pix-tensorflow
Nov 2016 - Sep 2019
TensorFlow implementation of "Image-to-Image Translation Using Conditional Adversarial Networks".
Role in this project:ML Engineer
ML EngineerContributions:35 commits, 4 PRs, 17 pushes in 2 years 9 months
Contributions summary:Yen contributed to the TensorFlow implementation of pix2pix, a conditional adversarial network for image-to-image translation. Their work includes fixing bugs in the model, adding test functionalities, updating the download dataset script, and modifying file paths for dataset access. The user focused on enhancing the testing process and ensuring proper image loading and pre-processing for both training and testing.
cyclegandeep-learningadversarialconditionaltranslation
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.