Yi Liu
Researcher at Pepole.cn
China
Summary
🤩
Rockstar🎓
Top SchoolYi Liu is a researcher and seasoned machine learning engineer with 10 years’ experience building computer vision and pattern recognition systems and production-grade applications. He holds a PhD in Computer Engineering from the Institute of Computing Technology, Chinese Academy of Sciences, and has driven ML platform and distributed-training features at major Chinese tech firms including Baidu and Tencent before joining Pepole.cn. Yi contributed directly to PaddlePaddle’s core distributed training stack—implementing and refactoring collective communication ops for LocalSGD and GradAllReduce—bridging research algorithms with scalable backend engineering. He combines deep academic training with hands-on backend and API work, routinely translating advanced optimization techniques into robust, deployable code. Based in China, he is particularly strong at turning novel CV research into practical, distributed implementations for real-world systems.
10 years of coding experience
9 years of employment as a software developer
Bachelor's degree, Computer Science, Bachelor's degree, Computer Science at Beihang UniversityDoctor of Philosophy - PhD, Computer Engineering, Doctor of Philosophy - PhD, Computer Engineering at Institute of Computing Technology, Chinese Academy of Sciences
Github Skills (11)
paddlepaddle10
deeplearning-ai10
machine-learning10
distributed-training10
deep-learning10
python10
scalability9
neural-network9
c-language8
cuda8
cprogramming-language8
Programming languages (6)
Github contributions (5)
PaddlePaddle/Paddle
Jul 2019 - Sep 2020
PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:41 reviews, 25 commits, 132 PRs in 1 year 2 months
Contributions summary:Yi's primary contribution focused on enhancing the distributed training capabilities of the PaddlePaddle framework. This involved implementing and refactoring collective communication operations for distributed training, specifically for LocalSGD and GradAllReduce optimizers. The changes included modifying the underlying collective communication infrastructure, adding new operators, and adapting the existing code to support the new features. Additionally, the user worked on improving the distributed classification training and updating API specifications.
pytorchpythonparalleldeep-learningpaddlepaddle
gavin1332/paddle-demo
Mar 2020 - Apr 2021
sample code for sharing
Contributions:75 commits, 1 PR, 91 pushes in 1 year 1 month
paddlepaddlesiggraphpaddlesharing
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.