Pengyun Wang
Research Assistant at Carnegie Mellon University
Pittsburgh, Pennsylvania, United States
Summary
🤩
Rockstar🎓
Top SchoolPengyun Wang is a research-focused engineer nearing completion of a PhD at Carnegie Mellon University with four years of applied experience at the intersection of engineering, data science, and machine learning. As a Research Assistant she models dynamic systems and studies transitions in injection-induced seismicity, applying advanced mathematical and statistical methods to real-world environmental problems. She has hands-on ML engineering experience contributing to OpenOOD, implementing DRAEM networks for anomaly and out-of-distribution detection and integrating them into benchmarking workflows. Based in Pittsburgh, Pengyun blends domain expertise in environmental engineering with practical deep learning skills, making her well-suited for roles that require translating complex research into deployable models and robust evaluation pipelines.
4 years of coding experience
Bachelor's degree, Environmental Engineering, Bachelor's degree, Environmental Engineering at Sun Yet-sen UniversityMaster of Science (M.S.), Environmental Engineering, Master of Science (M.S.), Environmental Engineering at CMU
Chinese, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (7)
neural-network10
pytorch10
convolutional-neural-networks10
deep-learning10
anomaly-detection10
python9
machine-learning9
Programming languages (2)
Github contributions (5)
Jingkang50/OpenOOD
Jan 2022 - Oct 2022
Benchmarking Generalized Out-of-Distribution Detection
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 73 commits, 23 PRs in 8 months
Contributions summary:Pengyun contributed code related to the DRAEM (Deep Robust Anomaly Extraction Model) for out-of-distribution detection within the context of the openood project. Their commits involved the implementation of a deep learning model, specifically the `DRAEM_networks.py` file, which included classes for Reconstructive and Discriminative SubNetworks. The user introduced features for anomaly detection and localization, as well as the configuration of the DRAEM model for training. They also updated the evaluator and dataset to support DRAEM integration within openood.
benchmarkingoutlier-detectionrobustnessanomaly-detectionbenchmark
Prophet-C/OpenOOD_local
Jun 2022 - Jan 2023
Benchmarking Generalized Out-of-Distribution Detection
Contributions:21 pushes in 6 months
benchmarkingbenchmarkout-of-distribution-detectionout-of-distributiondistribution
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.