Yuqiu Ji is a Senior LLM Engineer with 13 years of software engineering experience, blending deep machine learning model optimization with practical full-stack and tooling expertise. Based in Nantong, China, he has driven proprietary model alignment and self-improving instruction pipelines at Xiaohongshu that matched much larger baselines while reducing low-value training data by 63%. Previously at Meituan he led data-system architecture and training strategies, compressing instruction datasets to 10% of their original size without losing performance and owning a 33B base model release. An active open-source contributor and longtime front-end/build-system engineer, his commits to FIS/FIS3 and related projects show strong instincts for build reliability, cross-platform native addon packaging, and developer ergonomics. With a CS master from Harbin Institute of Technology and a linguistics BA from Fudan, he combines language-aware thinking with engineering rigor to optimize LLMs for both general and specialized domains.
12 years of coding experience
3 years of employment as a software developer
Bachelor's degree, Linguistics, Bachelor's degree, Linguistics at Fudan University
Master's degree, Computer Science, Master's degree, Computer Science at Harbin Institute of Technology
Contributions:134 commits, 98 pushes, 1 branch in 3 years 7 months
Contributions summary:Yuqiu contributed to the front-end and build process of the FIS-plus project, a front-end integration solution. They made updates to the core `fis-plus.js` file, including setting default values and charset encoding. They also enhanced the build process by implementing automatic generation of `smarty.conf` and integrating a livereload feature for development.
Cross-platform image decoder(png/jpeg/gif) and encoder(png/jpeg) for Nodejs
Role in this project:
Back-end Developer
Contributions:3 releases, 1 review, 122 commits in 8 years 8 months
Contributions summary:Yuqiu primarily focused on dependency management and build configuration for the node-images project. They updated dependencies, adjusted the build process, and added binary packages for different operating systems and architectures. Their work involved modifying build files, shell scripts for library checks, and JavaScript files to ensure the correct native addon is loaded for various Node.js versions. These changes were critical for the project's cross-platform compatibility and successful deployment of native image processing capabilities.
jpegapngimagepng-decoderpsd
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.