Wang Dong is a founding engineer and engineering leader at Kumo.ai in Atherton, California, with nine years of experience building production ML, search and cloud systems. He has driven high-impact work at Pinterest, Wish, Nuro and Coupang — from building an in-house search engine and boosting search engagement to achieving 20x training throughput and massive latency and revenue improvements. He spans product, cloud infrastructure and model engineering, and contributes to open-source ML tooling such as PyTorch Geometric, where he implemented heterogeneous-graph support, lazy GNN initialization and loader enhancements. His GitHub bio and background hint at algorithmic trading and Topcoder competitive programming chops, bringing algorithmic rigor to system design. He holds degrees from Tsinghua and UCLA and now advises startups while scaling ML from research prototypes to production.
Contributions:39 reviews, 6 commits, 6 PRs in 1 year
Contributions summary:Dong primarily contributed to the PyTorch Geometric library, focusing on modifying and extending the HeteroData class, and implementing features related to heterogeneous graphs. Their work involved API modifications, addressing comments, and fixing unit tests, with a strong emphasis on graph data structures. The user also worked on lazy GNN initialization across different convolutional layers and the implementation of automatic `n_id` and `e_id` attributes for loaders, indicating a focus on both data handling and model development.
Contributions:8 pushes, 1 branch in 2 years 11 months
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.