Bryan He
Computer Scientist
Summary
🤩
RockstarBryan He is a computer scientist with 13 years of practical experience bridging research and applied machine learning, currently a PhD student in Computer Science at Stanford. He contributes to influential open-source ML tooling—most notably enhancing Snorkel’s generative modeling to better support priors, fixed weights, and supervised categorical learning—demonstrating a strength in weak supervision. His work spans time-series and video analysis using LSTM architectures and modern visualization techniques like UMAP and t-SNE, reflecting both modeling depth and attention to interpretability. Comfortable moving between research-grade code and production-focused engineering, he pairs rigorous diagnostics and testing with experimental feature engineering to improve model robustness.
12 years of coding experience
Github Skills (31)
pytorch10
python10
data-science10
snorkel10
scikit10
machine-learning10
recurrent-neural-networks10
numpy10
lstm10
image-segmentation10
segmentation10
scikit-learn10
neural-network10
computer-vision10
generative-model10
Programming languages (5)
Github contributions (5)
bryanhe/dynamic
Feb 2021 - Feb 2023
Role in this project:
Data Scientist / ML Engineer
Data Scientist / ML EngineerContributions:1 release, 343 commits, 274 pushes in 2 years
Contributions summary:Bryan focused on developing and refining a machine learning model for video analysis within the repository. Their work involved implementing and experimenting with LSTM-based recurrent neural networks for time series data. They also explored methods for feature extraction and analysis, including the use of UMAP and t-SNE for visualization and dimensionality reduction. The user's contributions appear centered on improving the model's performance and integrating segmentation techniques.
snorkel-team/snorkel
Aug 2016 - Jun 2017
A system for quickly generating training data with weak supervision
Role in this project:ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:54 commits, 3 PRs, 59 pushes in 9 months
Contributions summary:Bryan primarily focused on enhancing the generative modeling capabilities within the Snorkel framework. They made significant contributions to the `GenerativeModel` class, introducing and refining features related to priors, fixed weights, and supervised learning integration for categorical variables. Their work involved modifying the model's training process, diagnostics, and weight processing to accommodate different scenarios in weak supervision. The user also added tests to validate the new functionalities, ensuring accurate behavior.
weak-supervisionpythondata-sciencemachine-learninglabeling
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.