Xuanlei Zhao
Research Engineer at Adobe
Singapore
Summary
🤩
Rockstar🎓
Top SchoolXuanlei Zhao is a research engineer with seven years of experience specializing in machine learning systems and large-scale model optimization, currently based in Singapore and working at Adobe. With a PhD trajectory at NUS and hands-on roles at ColossalAI, he has contributed to making large AI models cheaper and faster—adding flash attention and scheduler optimizations to a widely used open-source library. He blends deep research training with practical engineering, shipping extra functionality from research projects into production-grade tooling. Xuanlei’s profile reflects a knack for performance-focused model engineering and an ability to translate novel research into code that scales.
7 years of coding experience
1 year of employment as a software developer
Bachelor of Engineering - BE, Computer Science & Electronic Information Engineering, Bachelor of Engineering - BE, Computer Science & Electronic Information Engineering at Huazhong University of Science and TechnologyDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at National University of Singapore
Github Skills (8)
pytorch10
machine-learning10
deep-learning10
python10
ai10
large-scale10
parallelization9
inference8
Programming languages (4)
Github contributions (5)
hpcaitech/ColossalAI
Jul 2022 - Jan 2023
Making large AI models cheaper, faster and more accessible
Role in this project:
ML Engineer
ML EngineerContributions:95 reviews, 141 commits, 106 PRs in 6 months
Contributions summary:Xuanlei primarily contributed to the development of the `colossalai` library, which focuses on large AI models. Their work included code style improvements within the learning rate scheduler and the addition of flash attention functionality, showcasing an interest in optimization for deep learning models. The user also implemented and integrated extra functionality related to a new research project in this library.
heterogeneous-trainingcolossal-aifinetuningdeep-learninginference
oahzxl/OpenDiT
Apr 2024 - Sep 2024
OpenDiT: An Easy, Fast and Memory-Efficient System for DiT Training and Inference
Contributions:37 PRs, 162 pushes, 42 branches in 4 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.