Jieyao Gao
Engineering Manager - Advertiser AI Agent at Instagram
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJieyao Gao is an engineering manager and machine learning practitioner with a decade of experience building high-performance recommender and compiler systems for large-scale platforms. Currently leading Advertiser AI Agent efforts at Meta in Mountain View, Jieyao previously held progressive engineering and leadership roles across Instagram and Meta, blending hands-on optimization with team-level delivery. His open-source contributions to PyTorch projects—torchrec, Glow, and FBGEMM—reflect deep expertise in sparse architectures, quantized embeddings, kernel parallelism, and CPU/AVX optimization that drive faster inference at scale. Trained in computer science (UConn) with roots in communications engineering, he brings a rare combination of low-level performance engineering and product-focused ML systems leadership. Notably, he has implemented multithreading for sparse GPU inference and integrated GELU into hardware compiler backends, showing a knack for turning research-grade techniques into production wins.
10 years of coding experience
6 years of employment as a software developer
Bachelor of Engineering (B.Eng.) Communication Engineering, Bachelor of Engineering (B.Eng.) Communication Engineering at University of Shanghai for Science and TechnologyMaster of Science (M.S.) Computer Science, Master of Science (M.S.) Computer Science at University of Connecticut
mandarian, English
Github Skills (31)
pytorch10
recommendation-system10
c-language10
fba10
emm10
avx10
machine-learning10
sharding10
deep-learning10
f10
gpu10
performance-optimization10
compile10
neural-network10
cuda10
Programming languages (3)
Github contributions (5)
pytorch/FBGEMM
Aug 2020 - Aug 2022
FB (Facebook) + GEMM (General Matrix-Matrix Multiplication) - https://code.fb.com/ml-applications/fbgemm/
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 review, 23 commits, 33 PRs in 2 years
Contributions summary:Jieyao's commits primarily focused on optimizing and enhancing the FBGEMM library, specifically concerning groupwise convolution. Their contributions involved refactoring code, adding support for AVX512 and AVX512 VNNI, and addressing bugs related to weight packing and CPU initialization. The user also made significant contributions to improving code reusability and architecture-specific numerics codegen, contributing to the performance and versatility of the library.
machine-learningmatrix-multiplication
meta-pytorch/torchrec
Jan 2022 - Sep 2022
Pytorch domain library for recommendation systems
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:25 commits, 57 PRs in 8 months
Contributions summary:Jieyao primarily focused on enhancing the performance and capabilities of the PyTorch-based recommendation system library. Their contributions included implementing multi-threading for sparse architectures to accelerate GPU inference, specifically within the `torchrec` framework. They also worked on integrating and optimizing sequence embedding architectures, including support for quantized embeddings, to improve inference efficiency and support variable batch sizes. The user's work encompassed both back-end optimizations and the integration of advanced machine learning techniques.
pytorchrecommendation-systemgpudeep-learningcuda
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.