Zhaocheng Zhu
Senior Research Scientist at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolZhaocheng Zhu is a Senior Research Scientist at NVIDIA with nine years of experience spanning LLM post-training, alignment, reasoning, and machine learning systems. He holds advanced degrees from Mila and an undergraduate degree from Peking University, and has conducted research at Google and internships at Microsoft and Mitsubishi Electric, reflecting deep academic and industry cross-pollination. His open-source contributions to graph learning projects—improving core GPU code, compatibility, and documentation in widely used repos like GraphVite and TorchDrug—showcase a practical focus on robustness and performance. Zhaocheng blends rigorous research with hands-on engineering, often surfacing in maintenance, tooling, and reproducibility work that quietly accelerates broader ML progress. Based in California, he brings both theory-driven insight and production-minded fixes to large-scale model training and deployment challenges.
9 years of coding experience
1 year of employment as a software developer
Doctor of Philosophy - PhD, Doctor of Philosophy - PhD at Mila - Quebec Artificial Intelligence InstituteBachelor's degree Computer Science, Bachelor's degree Computer Science at Peking University
Github Skills (20)
pytorch10
restructuredtext10
python10
machine-learning10
rs10
deep-learning10
knowledge-graph10
gpu10
cuda10
graph-neural-network10
documentation10
conda9
data-visualization8
apidoc8
cprogramming-language8
Programming languages (4)
Github contributions (5)
DeepGraphLearning/torchdrug
Aug 2021 - Nov 2022
A powerful and flexible machine learning platform for drug discovery
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:3 releases, 62 commits, 13 PRs in 1 year 2 months
Contributions summary:Zhaocheng primarily contributed to fixing compatibility issues with PyTorch versions and ensuring CPU compatibility within the `torchdrug` platform. They addressed the pinned Python version in Conda environments and resolved metaclass issues in `patch.py`. Furthermore, the user added new variadic functions such as `variadic_sample`, `variadic_meshgrid`, and `variadic_to_padded` for enhanced functionality. They also implemented improvements to the installation documentation, enhanced the overall performance of the spmm/rspmm functionality, and included fixes for a bug related to reloading scheduler configurations.
pytorchmachine-learning-platformlearning-platformdataminingdrug-discovery
DeepGraphLearning/graphvite
Aug 2019 - Jan 2021
GraphVite: A General and High-performance Graph Embedding System
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:15 commits, 32 pushes, 1 branch in 1 year 5 months
Contributions summary:Zhaocheng primarily focused on fixing and improving core components within the GraphVite system. Their contributions included addressing knowledge graph initialization issues and modifying index types in the GPU code. The user also worked on the command-line interface for visualization. These changes indicate a focus on improving the functionality and stability of the graph embedding system.
graph-neural-networksperformancegraphgraph-embeddingembedding
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.