Tues Tang
Freelance at Lucky Global Services
Monterey Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolTues Tang is a versatile freelance engineer with 11 years of experience bridging practical business roles and hands-on technical contributions, currently based in Monterey Park, California. With a background spanning accounting, content management, and fashion/business studies, Tues brings a pragmatic, multidisciplinary approach to problem solving and teamwork. As a PhD student at Tsinghua and an active contributor to the high-profile Apache TVM project, they have implemented CUDA backend optimizations—improving cross-thread reductions and warp memory handling—to boost deep learning performance. Comfortable working solo or in teams, Tues combines a hardworking yet efficient work style with a knack for unconventional thinking that finds elegant, performance-oriented solutions.
11 years of coding experience
DNIIT, .NET Programming, DNIIT, .NET Programming at NIITFashion, Business Management, Fashion, Business Management at PCCAssociate's degree, Accounting, Associate's degree, Accounting at Lotus University
Chinese, Vietnamese, English, Chinese
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (17)
tvm10
performance-monitor10
machine-learning10
deeplearning-ai10
compiler-compiler10
deep-learning10
performance-analysis10
gpu10
cuda10
compiler10
python9
vulkan8
opencl8
cprogramming-language8
roc8
Programming languages (17)
Github contributions (5)
apache/tvm
Apr 2020 - Sep 2020
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
ML Engineer & Performance Engineer
ML Engineer & Performance EngineerContributions:5 reviews, 10 commits, 19 PRs in 5 months
Contributions summary:Tues contributed to the development and optimization of the TVM compiler stack, specifically focusing on integrating cross-thread reduction and improving warp memory usage within the CUDA backend. They implemented a mix of normal and cross-thread reductions, improving the performance of deep learning computations. Furthermore, the user improved warp memory management and fixed codegen for warp shuffle intrinsics to enhance the efficiency of the compiler. Additionally, the user fixed a typo and improved documentation.
metalvulkancompilertensoropencl
chyyuu/os2018-roastduck
Mar 2018 - May 2018
Contributions:77 commits in 2 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.