Xiaoyu Zhai
Senior Machine Learning Engineer at Kuaishou Technology
Beijing, China
Summary
👤
Senior🎓
Top SchoolXiaoyu Zhai is a Senior Machine Learning Engineer with 11 years of experience building scalable data infrastructure and production ML systems, currently accelerating transformer-like and LLM inference and content-moderation pipelines at Kuaishou in Beijing. Proficient in Python, Golang, Kubernetes, Docker, Ray, and Linux, Xiaoyu blends low-level backend robustness with high-performance distributed training and serving expertise. He contributes to notable open-source projects like DGL-KE, improving distributed training reliability and data handling for large-scale knowledge graph embeddings. A Stevens Institute of Technology alumnus with a BE from Sichuan University, he focuses on pragmatic, production-first solutions that bridge research models and real-time serving constraints. Not obvious from his title: he often tackles gritty infrastructure edge cases—network interface handling, partition path resilience, and API robustness—that keep large distributed workloads running smoothly.
11 years of coding experience
Bachelor of Engineering (BE), Bachelor of Engineering (BE) at Sichuan UniversityMaster of Science (MS), Master of Science (MS) at Stevens Institute of Technology
Chinese, English
Stackoverflow
103reputation
3kreached
2answers
3questions
Github Skills (13)
machine-learning10
python10
dgl10
distributed-training9
error-handling8
file-management7
roslyn6
roslyn-code-analysis6
electron6
glsl6
electron-rebuild6
geometry-shader6
ffi6
Programming languages (8)
Github contributions (5)
awslabs/dgl-ke
Apr 2021 - Sep 2021
High performance, easy-to-use, and scalable package for learning large-scale knowledge graph embeddings.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:5 commits, 6 PRs, 8 comments in 5 months
Contributions summary:Xiaoyu contributed to several aspects of the DGL-KE project, focusing on code enhancements and bug fixes. Their work includes addressing network interface availability issues, enabling recursive directory creation for partition paths, and integrating a data_files argument for custom datasets. Additionally, they fixed a missing positional argument in a function call and added checks for the existence of attributes before using them. These changes indicate a focus on improving the robustness and functionality of the distributed training components.
graph-learningknowledge-graphs-embeddingsgraph-embeddingsknowledge-graphlarge-scale
ryantd/Movie-Recommendation-Web-App
Apr 2016 - Feb 2017
Contributions:20 commits, 30 pushes, 4 branches in 10 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.