Yang Xue
Graduate Teaching Assistant
Shanghai, China
Summary
🤩
Rockstar🎓
Top SchoolYang Xue is an ML-focused physicist and graduate teaching assistant with nine years of hands-on experience developing and refining rotated object detection systems in TensorFlow and the OpenMMLab ecosystem. His open-source contributions include core implementations and optimizations to R3Det and mmrotate, adding algorithmic support (CSL, ATSS, Stable KLD) and practical dataset integrations like DOTA1.5 and DIOR-R. Combining experimental lab automation experience from UC Berkeley with analytic work on topological responses at ShanghaiTech, he bridges theoretical physics and applied ML tooling. Based in Shanghai and pursuing a PhD at The Ohio State University, he brings a research-first mindset to engineering problems and routinely improves model training, configs, and test pipelines. Colleagues value his knack for turning nuanced algorithmic ideas into robust code and reproducible benchmarks.
9 years of coding experience
2 years of employment as a software developer
Bachelor of Applied Science - BASc, Physics, Bachelor of Applied Science - BASc, Physics at ShanghaiTech UniversityPhysics, Physics at University of California, BerkeleyDoctor of Philosophy, Physics, Doctor of Philosophy, Physics at The Ohio State University
Github Skills (22)
algorithm10
algorithms10
pytorch10
python10
machine-learning10
detectors10
data-structure10
detector10
rotation10
trainings10
tensorflow10
retinanet10
data-structures10
object-detection10
computer-vision10
Programming languages (4)
Github contributions (5)
yangxue0827/RotationDetection
Oct 2020 - Jan 2023
This is a tensorflow-based rotation detection benchmark, also called AlphaRotate.
Role in this project:
ML Engineer
ML EngineerContributions:3 releases, 160 commits, 2 PRs in 2 years 3 months
Contributions summary:Yang implemented support for the RetinaNet and R3Det models by modifying the code to incorporate the new models. They integrated support for new datasets, specifically DOTA1.5 and DIOR-R, and added new features to the existing codebase, such as the KLD loss. The contributions indicate active development and extension of the rotation detection benchmark.
tensorflow
Thinklab-SJTU/R3Det_Tensorflow
Nov 2019 - Jun 2021
Code for AAAI 2021 paper: R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object
Role in this project:ML Engineer
ML EngineerContributions:63 commits, 2 PRs, 53 pushes in 1 year 7 months
Contributions summary:Yang primarily contributed to updating and training baseline weights, and adding core code related to R3Det, a refined single-stage detector for rotating objects. Their work involved modifications to configuration files, suggesting involvement in model training and optimization within the TensorFlow framework. These changes indicate a focus on improving object detection performance, likely through refinement of the model's architecture or training process. The commits also show the user's engagement in model testing by updating test scripts.
pytorchrotationrefinedrefinementrotating
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.