Yuan Liu
Algorithm Researcher at Shanghai AI Laboratory
Minhang District, Shanghai, China
Summary
🤩
Rockstar🎓
Top SchoolYuan Liu is an algorithm researcher with 8 years of experience specializing in computer vision and multimodal AI, currently working at Shanghai AI Laboratory. He has practical experience shipping robust image processing pipelines and training infrastructure, contributing data transform fixes and caching to the widely used OpenMMLab/mmcv and core hook architecture to mmengine. Yuan also integrated and adapted leading multimodal models (MiniGPT-4, InstructBLIP, OpenFlamingo) into the OpenCompass LLM evaluation platform, demonstrating skill bridging research models and evaluation tooling. His background includes internships at ByteDance and Horizon Robotics, and a Master’s in Electrical and Electronics Engineering from the University of Hong Kong. Known on GitHub as an "AI monster," he pairs hands-on engineering with documentation and DevOps improvements, including Dockerized documentation builds for mmselfsup. Notably, he focuses on making research code production-friendly through refactors, compatibility fixes, and rigorous testing.
8 years of coding experience
2 years of employment as a software developer
Bachelor's Degree, Electrical and Electronics Engineering, Bachelor's Degree, Electrical and Electronics Engineering at Southeast UniversityThe University of Hong Kong (HKU)
Github Skills (31)
markdown10
pytorch10
hook10
docker10
restructuredtext10
python10
image-processing10
sphinx10
openmmlab10
machine-learning10
rs10
dockers10
numpy10
llm10
deep-learning10
Programming languages (3)
Github contributions (5)
open-mmlab/mmselfsup
Dec 2021 - Dec 2022
OpenMMLab Self-Supervised Learning Toolbox and Benchmark
Role in this project:
Technical Writer & DevOps Engineer
Technical Writer & DevOps EngineerContributions:1 release, 339 reviews, 92 commits in 1 year
Contributions summary:Yuan's commits primarily focus on enhancing the project's documentation and setting up the necessary infrastructure for building and deploying the project. They added documentation files and configured the documentation build process. The user refactored the directory structure related to documentation, fixed links within the documentation, and refined the appearance of documentation elements. Their contributions also included setting up and modifying Docker configurations for building the project.
pytorchsupervised-learningbeitsimclrmae
open-compass/opencompass
Jul 2023 - Nov 2023
OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.
Role in this project:ML Engineer
ML EngineerContributions:65 reviews, 30 PRs, 34 pushes in 4 months
Contributions summary:Yuan primarily focused on integrating and adapting various multimodal models within the OpenCompass LLM evaluation platform. Their contributions include adding support for MiniGPT-4, InstructBLIP, and OpenFlamingo, along with configurations for several public datasets. The user also refactored input/output processes and made updates to existing task configurations. Additionally, they addressed dependencies, fixing import errors, and ensuring compatibility with new versions of the models.
benchmarkchatgptevaluationlarge-language-modelllama2
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.