Yao Yao
Postdoctoral Researcher at Purdue University
Lafayette, Indiana, United States
Summary
👤
Senior🎓
Top SchoolYao Yao is a postdoctoral researcher and software engineer with 11 years of experience building data-processing and ML-driven solutions for the CMS experiment at CERN, now based at Purdue University. He combines hands-on detector work—assembling and testing CSC muon electronics—with advanced data analysis searching for dark photons and a prospective top–antitop bound state in CMS Run II data. Technically fluent in C++, Python, and shell scripting, he focuses on optimizing machine learning models and embedding inference servers (NVIDIA Triton) into high-throughput HEP workflows. As an active contributor to the widely used cms-sw/cmssw framework, he integrated a "SONIC-ized" Particle Transformer to accelerate B-tagging via heterogeneous computing. This blend of hardware, analysis, and production ML engineering gives him a rare end-to-end perspective on turning collider data into deployable, performant inference pipelines.
11 years of coding experience
Doctor of Philosophy - PhD, Physics, Doctor of Philosophy - PhD, Physics at University of California, Davis
Spanish
Github Skills (12)
heterogeneous10
machine-learning10
c-language10
onnx10
cprogramming-language10
particle-physics9
cplus9
cms9
cpp9
cer9
triton8
python4
Programming languages (4)
Github contributions (5)
cms-sw/cmssw
Jun 2021 - Jun 2021
CMS Offline Software
Role in this project:
ML Engineer
ML EngineerContributions:3 reviews, 1 commit, 4 PRs in 1 day
Contributions summary:Yao primarily focused on integrating and optimizing a Particle Transformer model within the CMS offline software framework. Their contributions include adding a "SONIC-ize" version of the Particle TransformerAK4, modifying configuration files to incorporate the new model, and moving common inputs to the interface and source files. This work suggests an effort to leverage heterogeneous computing (e.g. Triton) to improve performance for the B-tagging algorithm in the context of the CMS experiment. The user also responded to code comments and updated the format.
cmscernweb-appc-plus-plusbackbonejs
y19y19/GitHubTest
Jun 2022 - Oct 2024
Contributions:8 pushes, 2 branches in 2 years 4 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.