Yutong Zhao
Principal Product Manager at NVIDIA
Cambridge, Massachusetts, United States
Summary
👤
Senior🎓
Top SchoolYutong Zhao is a Principal Product Manager with 13 years of experience at the intersection of computational chemistry, machine learning, and high-performance software, now based in Cambridge, MA and recently joining NVIDIA. Previously a Senior Director at Relay Therapeutics and senior engineering roles at Schrödinger and Stanford, she combines deep domain knowledge in chemistry with hands-on contributions to open-source scientific tools like DeepChem and OpenMM. Her work spans data pipelines for unlabeled molecular datasets, performance and CUDA fixes for molecular simulation kernels, and serialization support—showing both research rigor and production-grade engineering. Known for translating complex simulation and ML needs into robust product and engineering outcomes, she brings a rare mix of scientific training (MPhil in CS & Chemistry) and practical implementation experience.
13 years of coding experience
13 years of employment as a software developer
HBSc meth & chemistry, HBSc meth & chemistry at University of TorontoHong Kong University of Science and Technology (HKUST)
Github Skills (25)
simulation10
python10
molecular-simulation10
data-science10
splits10
simulations10
machine-learning10
data-transformation10
data-loading10
deep-learning10
molecular-dynamics-simulation10
nonlinear-dynamics10
cuda10
splitting10
pandas9
Programming languages (11)
Github contributions (5)
openmm/openmm
Sep 2012 - Jan 2019
OpenMM is a toolkit for molecular simulation using high performance GPU code.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:49 commits, 1 PR, 32 comments in 6 years 5 months
Contributions summary:Yutong focused on fixing padding errors and CUDA-memcheck errors within the `CudaCalcNonbondedForceKernel`, addressing issues related to nonbonded force calculations in the CUDA platform. They also worked on implementing FFTWs to work with complex conjugate arrays, and decoupling energy and force calculations into separate kernels. Furthermore, the user made contributions to fixing potential energy discrepancies in CUDA and Reference platform, involving the energy calculation method. Additional contributions include support for serialization and deserialization in both Python and the base class in the codebase.
cudamolecular-dynamicsopenmmgpusimulation
deepchem/deepchem
Jan 2017 - Nov 2017
Democratizing Deep-Learning for Drug Discovery, Quantum Chemistry, Materials Science and Biology
Role in this project:Back-end Developer & Data Scientist
Back-end Developer & Data ScientistContributions:41 commits, 20 PRs, 235 comments in 9 months
Contributions summary:Yutong contributed significantly to the `deepchem` repository, focusing on supporting unlabelled datasets. They implemented and tested features for loading and transforming unlabelled datasets within the data loading and transformation pipelines. The user also added tests for these new functionalities. In addition, the user integrated ButinaSplitter class and added support for bench marking the tox21 dataset. The code changes involved modifications to data loading, data transformations, and data splitting, demonstrating proficiency in working with datasets and machine learning workflows.
chemistrysciencedrug-discoverydeep-learningbiology
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.