Yuqing Liu
Research And Development Engineer at Baidu Inc.
Shanghai, Shanghai, China
Summary
🤩
Rockstar🎓
Top School
Top expert inArtificial Intelligence and Computer Vision Technologies
Yuqing Liu is a Research and Development Engineer at Baidu with nine years of experience specializing in backend performance engineering for deep learning frameworks and edge inference. Based in Shanghai, she has made significant open-source contributions to the PaddlePaddle ecosystem—optimizing Paddle-Lite for x86, improving core operators and execution flow, and enhancing model training and benchmarking for CV and OCR tasks. Her work spans low-level operator optimizations, runtime efficiency improvements, and practical tooling like profiling and multi-process data loaders, reflecting a blend of systems thinking and ML engineering. She also contributes to documentation, bridging developer experience with implementation details, and has a knack for squeezing CPU performance (e.g., avoiding unnecessary copies in loss computations). Colleagues rely on her to translate algorithmic needs into performant, production-ready code that powers both research and deployment.
9 years of coding experience
Huazhong University of Science and Technology
Github Skills (29)
c-language10
python10
operation10
tensorrt10
ocr10
machine-learning10
deep-q-learning10
rnn-model10
n10
api-documentation10
deep-learning10
tensorflow10
performance-optimization10
optical-character-recognition10
performance-tuning10
Programming languages (7)
Github contributions (5)
PaddlePaddle/benchmark
Mar 2019 - Dec 2022
Role in this project:
ML Engineer
ML EngineerContributions:241 reviews, 243 commits, 580 PRs in 3 years 10 months
Contributions summary:Yuqing added a benchmark for an OCR attention model, including running scripts, configuration files, and training scripts. The user modified the code to implement the OCR attention model, primarily working on defining the encoder and decoder networks with attention mechanisms. They also optimized the infer program. The user also refined the timing of PaddingRNN, adding profiling capabilities and saving the inference model.
PaddlePaddle/Paddle
Dec 2017 - Jan 2023
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 & Performance Engineer
Back-end Developer & Performance EngineerContributions:1274 reviews, 237 commits, 854 PRs in 5 years 2 months
Contributions summary:Yuqing's commits primarily focused on optimizing and refining the PaddlePaddle framework's execution engine. Contributions include refining argument orders, merging code changes, and correcting typos within the `executor` and `framework` modules. The user also optimized the performance of `linspace`, improved the implementation of `bce_loss` to avoid CPU copies, and implemented features for the `fusion_group` operator.
pytorchpythonparalleldeep-learningpaddlepaddle
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.