Tsung-yi Lin
Research Director at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolTsung-yi Lin is a research leader with 13 years of experience in computer vision and machine learning, currently serving as Research Director at NVIDIA and Secretary of the Common Visual Data Foundation. He earned a PhD from Cornell Tech and has held senior research roles at Google, with internships at Facebook and Microsoft, reflecting deep industry-academic breadth. Tsung-yi’s work spans foundational dataset tooling and model research—he’s an active contributor to the widely used COCO API (pycocotools), improving efficiency and keypoint/mask handling for large-scale vision benchmarks. Known for translating rigorous research into production-ready libraries and systems, he combines algorithmic insight with practical engineering. Based in California, he brings a global perspective from Taiwanese and US education and early research experience, often focusing on making large vision datasets and tools more usable and performant.
13 years of coding experience
5 years of employment as a software developer
University of California, San DiegoBS, Electrical Engineering, BS, Electrical Engineering at National Taiwan UniversityDoctor of Philosophy (Ph.D.), Computer Vision & Machine Learning, Doctor of Philosophy (Ph.D.), Computer Vision & Machine Learning at Cornell University
Github Skills (8)
lib10
dataannotations10
python10
data-annotation10
image-processing9
performance-optimization9
feature-detection8
keypoint8
Programming languages (5)
Github contributions (5)
cocodataset/cocoapi
Jan 2015 - Feb 2020
COCO API - Dataset @ http://cocodataset.org/
Role in this project:
Back-end Developer
Back-end DeveloperContributions:73 commits, 10 PRs, 68 pushes in 5 years 1 month
Contributions summary:Tsung-yi primarily contributed to the development and improvement of the pycocotools library. The commits focused on adding new functionalities like `loadRes()` for result loading and keypoint support, as well as fixing bugs related to mask handling within the annotation loading process. The user also refactored and optimized the code, resulting in significant performance improvements, demonstrating a focus on library efficiency and usability for the Microsoft COCO dataset.
apidataset-apidatasetcocococodataset
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.