Yixuan He
Research Scientist at Arizona State University
United States
Summary
🤩
RockstarYixuan He is an Assistant Professor and machine learning researcher with seven years of experience bridging academic research and industry practice in graph neural networks and spatiotemporal modeling. He holds a DPhil in Statistics from the University of Oxford and spent three years as an Applied Scientist at AWS, where he translated research ideas into production-facing solutions. His open-source contributions include accuracy fixes and new model integrations for the widely used PyTorch Geometric ecosystem and implementing ASTGCN/MSTGCN spatiotemporal models that improve time-series graph processing. Earlier roles include building behavior-based anti-cheating systems at NetEase Games and delivering Python training at the University of Edinburgh, demonstrating a blend of applied engineering, pedagogy, and product-minded research. Known for attention to rigorous evaluation (e.g., fixing AUC calculation bugs) and reproducible examples, he emphasizes both model correctness and usability. He combines deep statistical training with hands-on ML engineering to move novel graph and temporal methods from prototype to real-world use.
7 years of coding experience
Github Skills (10)
graph10
pytorch10
machine-learning10
deep-learning10
graph-neural-network10
python10
gnn10
evaluation9
eval9
time-series-forecasting9
Programming languages (5)
Github contributions (5)
benedekrozemberczki/pytorch_geometric_temporal
Mar 2021 - Apr 2022
PyTorch Geometric Temporal: Spatiotemporal Signal Processing with Neural Machine Learning Models (CIKM 2021)
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 6 reviews, 156 commits in 1 year
Contributions summary:Yixuan contributed significantly to the implementation of the ASTGCN and MSTGCN models, which are likely related to spatiotemporal data processing. Their work involved adding the MSTGCN model, refining ChebConv implementations, and including example usage for both models, along with making documentation and typo fixes. The user demonstrated a strong understanding of graph neural network architectures and their application to time-series data, improving the functionality of the project.
machine-learning-modelspytorch-geometricsignal-processingpytorchgraph-neural-networks
pyg-team/pytorch_geometric
Mar 2021 - Oct 2021
Graph Neural Network Library for PyTorch
Role in this project:ML Engineer
ML EngineerContributions:8 commits, 3 PRs in 7 months
Contributions summary:Yixuan primarily contributed to improving the accuracy of the graph neural network models in the repository. They fixed a bug related to the calculation of the Area Under the Curve (AUC) in the `SignedGCN` model, updating the code to use positive link probabilities. Furthermore, the user integrated new features and datasets from external repositories, and made minor corrections to improve code readability.
graph-neural-networkpytorchgeometric-deep-learninggraph-neural-networksdeep-learning
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.