Wenxuan Peng
PHD Student at NTU-MMLab
City of Ithaca, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolWenxuan Peng is a first-year PhD student in Computer Science at Cornell University with four years of hands-on research and engineering experience in computer vision and robustness. Previously a Robotics Institute Summer Scholar at CMU and an algorithm intern at NTU-MMLab, Wenxuan has also built production-facing vision features during an internship at ByteDance. His open-source contributions to the OpenOOD benchmark highlight practical work on out-of-distribution detection, adding data augmentations like PixMix, CutMix training, and temperature-scaling postprocessors to improve model uncertainty and robustness. Based in Ithaca, he blends academic rigor with applied ML engineering and a track record of shipping experiments that bridge research and real-world systems.
4 years of coding experience
Bachelor candidate, Computer Science, Year 3, Bachelor candidate, Computer Science, Year 3 at Nanyang Technological UniversityDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Cornell UniversityBachelor of Engineering - BE, Computer Science, Exchange, Bachelor of Engineering - BE, Computer Science, Exchange at National University of Singapore
Github Skills (6)
pytorch10
data-augmentation10
modeling9
python9
trainings9
machine-learning9
Programming languages (4)
Github contributions (5)
Jingkang50/OpenOOD
May 2022 - May 2022
Benchmarking Generalized Out-of-Distribution Detection
Role in this project:
ML Engineer
ML EngineerContributions:31 commits, 11 PRs, 1 push in 15 days
Contributions summary:Wenxuan primarily contributes to the `openood` repository, which focuses on out-of-distribution detection. The commits show additions of preprocessor structures, specifically related to PixMix, which is a data augmentation technique. Moreover, the user added a CutMix trainer and a temperature scaling postprocessor. The user's work indicates a focus on improving model robustness and uncertainty estimation through the application of different training techniques and post-processing strategies.
benchmarkingoutlier-detectionrobustnessanomaly-detectionbenchmark
LilyDaytoy/tp
Sep 2021 - Nov 2021
Contributions:2 PRs, 67 pushes, 4 branches in 1 month
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.