Sangdoo Yun
Seoul, Korea, Republic of
Summary
🤩
RockstarSeulki Park is a postdoctoral researcher in CSE at the University of Michigan with 10 years of experience building robust vision-language and structure-aware models that tackle real-world issues like data imbalance, noisy labels, and missing information. She combines strong academic training from Seoul National University with hands-on industry experience at NAVER and Samsung, publishing at CVPR and translating research into practical systems such as scalable aerial animal recognition and a lightweight multimodal sleep-staging model that outperformed commercial wearables. Her work emphasizes spatial and semantic hierarchies to boost efficiency and generalization, and she has a track record of improving underrepresented-class performance with low-cost augmentations. Open to research and applied scientist roles, she is motivated by deploying AI that remains reliable across imperfect, multimodal datasets and messy real-world taxonomies.
11 years of coding experience
Github Skills (40)
imagenet10
stylegan10
optimizer10
regularization10
mobilenet10
resnet10
transfer-learning10
efficientnet10
regularizer10
image-classification9
distributed-training9
pytorch9
clip9
loss9
deep-reinforcement-learning9
Programming languages (6)
Github contributions (5)
sangdooyun/sangdooyun.github.io
Jul 2019 - Jun 2026
Contributions:81 pushes in 6 years 11 months
clovaai/CutMix-PyTorch
Jul 2019 - Sep 2020
Official Pytorch implementation of CutMix regularizer
Contributions:9 commits, 10 pushes, 61 comments in 1 year 2 months
pytorchregularizercutmixaugmentationregularization
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.