Jiashuo Liu
Research Scientist at ByteDance
Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolJiashuo Liu is a research scientist with nine years of experience specializing in LLM data, evaluation, and robust ML systems, currently leading LLM auto-evaluation and pretraining work at ByteDance's Seed Team in Beijing. He holds a PhD from Tsinghua and has broad academic experience as a visiting researcher at Cambridge, Columbia Business School, and Stanford, where he focused on data-centric ML, robust optimization, and distributional robustness. His background blends rigorous theory and systems engineering—evident from contributions to Apache Pegasus where he improved backup, replication, and metadata resolution features for a high-performance key-value store. Comfortable moving between research and production, he has applied ML to healthcare, recommendation systems, and model stability evaluation, and brings a knack for making complex evaluation pipelines dependable and scalable.
9 years of coding experience
Doctor of Philosophy - PhD, Computer Science, GPA: 3.97/4.0 Rank: 3/129, Doctor of Philosophy - PhD, Computer Science, GPA: 3.97/4.0 Rank: 3/129 at 清华大学Visiting Student Researcher, Management Science and Engineering, Visiting Student Researcher, Management Science and Engineering at Stanford University
Github Skills (13)
sh10
script10
c-language10
shell10
cprogramming-language10
scripting10
thrift9
configuration-management9
distributed-systems8
system-design8
nosql7
distributed-database7
key-value-store7
Programming languages (11)
Github contributions (5)
apache/incubator-pegasus
Feb 2019 - Aug 2022
Apache Pegasus - A horizontally scalable, strongly consistent and high-performance key-value store
Role in this project:
Back-end Developer
Back-end DeveloperContributions:502 reviews, 289 commits, 104 PRs in 3 years 6 months
Contributions summary:Jiashuo primarily contributed to refactoring and extending the shell command implementations, particularly those related to backup policies, table environments, and general replication configuration. Their work involved modifying existing code, adding new APIs, and incorporating performance statistics, demonstrating a focus on system maintainability and feature enhancement. Furthermore, the user added support for resolving metadata through DNS and various improvements in the codebase.
scalabledbmsnosqlapacheconsistent
Shuo-Jia/pegasus-spark
Nov 2019 - Jul 2021
Contributions:217 pushes, 45 branches in 1 year 8 months
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.