Sheng Li
Engineering Manager at Baidu Inc.
Haidian District, Beijing, China
Summary
🤩
RockstarSheng Li is an experienced engineering manager with over a decade building and scaling QA and backend teams for cloud databases and storage systems, currently leading CloudDB quality and performance efforts at Baidu. He combines deep systems engineering roots—storage, Unix kernel, SAN/NAS architecture—with hands-on backend and ML framework contributions to projects like PaddlePaddle and PaddleNLP, where he implemented core tensor operations and optimized GPT-3 workflows. At Baidu he grew his team to ~30 engineers, owning regression, system, and performance testing while driving DevOps automation and competence building across the organization. Prior roles in storage R&D and project management give him a rare mix of product delivery discipline and low-level technical expertise. Colleagues know him for translating complex distributed-system risks into practical test standards and for cultivating career paths that retain technical talent.
9 years of coding experience
10 years of employment as a software developer
Github Skills (26)
transformers10
unit-testing10
distributed-training10
parallelization10
python10
machine-learning-models10
operation10
pre-trained-model10
model-driven10
tensorrt10
pretrain10
machine-learning10
llm10
model-building10
api-documentation10
Programming languages (4)
Github contributions (5)
PaddlePaddle/PaddleFleetX
Sep 2019 - Nov 2022
飞桨大模型开发套件，提供大语言模型、跨模态大模型、生物计算大模型等领域的全流程开发工具链。
Role in this project:
ML Engineer
ML EngineerContributions:7 releases, 154 reviews, 96 commits in 3 years 2 months
Contributions summary:Sheng primarily contributed to the development and integration of GPT models within the PaddleFleetX framework. Their work involved adding and modifying GPT model implementations, including a full GPT model and also a version for 3D parallelism. They introduced a pretraining example, added components, and implemented save/load functionalities for the GPT model.
resnetberttrainingbenchmarkpaddle
PaddlePaddle/Paddle
Aug 2019 - Dec 2022
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:666 reviews, 166 commits, 813 PRs in 3 years 4 months
Contributions summary:Sheng implemented the `eye` operation, including its kernel and unit tests, indicating a focus on core framework functionality. They also added the `gather_nd` and `scatter_nd` operations, along with their corresponding unit tests and error handling, reflecting their contribution to expanding the library's capabilities. Furthermore, the user worked on fixing documentation, coverage, and memory management, revealing involvement in framework optimization and maintenance.
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.