Hongyi Dong
Research Assistant at Penn State College of Information Sciences and Technology
State College, Pennsylvania, United States
Summary
👤
Senior🎓
Top SchoolHongyi Dong is a research-focused technologist with eight years of experience exploring how technology can empower youth agency and autonomy, currently working as a Research Assistant at Penn State’s Living Privacy Research Group. With an MLIS from UNC Chapel Hill and a BA from the University of Maine, Hongyi blends information science, privacy literacy, and hands-on machine learning engineering. On GitHub they’ve contributed substantive improvements to PyTorch object detection pipelines—optimizing Faster R-CNN and SSD training, dataset loading, and loss functions—which underscores a practical skillset in applied deep learning. Based in State College, Pennsylvania, they marry academic research sensibilities with production-minded code contributions and an ethos of curiosity (“too young, too simple, sometimes naive”) that drives experimentation. Colleagues can expect a collaborator who moves between privacy research and model engineering, bringing thoughtful design for empowering young users alongside solid implementation chops.
8 years of coding experience
Bachelor of Arts (BA), Bachelor of Arts (BA) at University of MaineMaster of Library & Information Science - MLIS, Master of Library & Information Science - MLIS at University of North Carolina at Chapel Hill
Japanese, Spanish, Tagalog, Chinese
Github Skills (12)
object-detection10
computer-vision10
pytorch10
faster-rcnn10
python10
modeling9
trainings9
ssds9
machine-learning8
vggnet8
resnet7
tensorboard7
Programming languages (3)
Github contributions (5)
dongdonghy/Detection-PyTorch-Notebook
Nov 2018 - Mar 2020
代码 -《深度学习之PyTorch物体检测实战》
Role in this project:
ML Engineer
ML EngineerContributions:10 commits, 22 pushes, 1 branch in 1 year 3 months
Contributions summary:Hongyi primarily contributes to the project by modifying and updating the training and evaluation scripts, configurations, and model definitions related to object detection. Their work involves adjustments to dataset loading, model architectures (VGG16, ResNet), and the training process within the PyTorch framework. The commits indicate a focus on adapting and optimizing the training pipeline for Faster R-CNN and SSD models. Additional changes include script updates and modifications to the loss functions.
deep-learningpytorch
dongdonghy/realtime_lane_detection
Mar 2018 - Mar 2018
Contributions:12 commits, 10 pushes in 1 day
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.