Bangtian Liu
Member Of Technical Staff AI Compiler at AMD
Old Toronto, Ontario, Canada
Summary
🤩
Rockstar🎓
Top SchoolBangtian Liu is an AI compiler engineer with nine years of experience optimizing compilers and GPUs for high-performance numerical computing, currently a Member of Technical Staff at AMD. He holds advanced research training from University of Toronto and Rutgers and has driven compiler and program verification research, hierarchical matrix approximation, and HPC on heterogeneous platforms. His open-source contributions include backend work for the prominent IREE ML compiler—adding GPU-focused distribution patterns, vector gather tests, and MMA intrinsic utilities—showing a strong focus on low-level performance tuning for modern accelerators. Prior roles at Microsoft and Tsinghua sharpened his skills across compiler stacks, program verification, and parallel programming education. Colleagues rely on him for bridging academic rigor with production-grade optimizations that squeeze peak performance from GPU hardware.
9 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of TorontoMaster’s Degree Computer Science, Master’s Degree Computer Science at Tsinghua UniversityBachelor’s Degree Computer Science, Bachelor’s Degree Computer Science at Huazhong University of Science and TechnologyDoctor of Philosophy (Ph.D.) Computer Engineering, Doctor of Philosophy (Ph.D.) Computer Engineering at Rutgers University
Chinese, English
Github Skills (13)
compiler10
machine-learning10
compiler-compiler10
gpu10
mlr10
llvm10
vectorization9
cuda9
roc8
tensorflow7
pytorch6
vulkan6
onnx6
Programming languages (5)
Github contributions (5)
iree-org/iree
May 2024 - Apr 2025
A retargetable MLIR-based machine learning compiler and runtime toolkit.
Role in this project:
Backend Developer & ML Engineer
Backend Developer & ML EngineerContributions:50 reviews, 61 PRs, 40 pushes in 11 months
Contributions summary:Bangtian contributed to the IREE project by implementing and modifying code related to machine learning compilation and runtime. Their work focused on enabling and testing features related to workgroup reordering and vector distribution, particularly for GPU code generation, specifically within the LLVMGPU backend. The contributions involved adding distribution patterns and test files for vector gather operations, supporting multi-dimensional reduction with scalars, and incorporating support for the contraction operation. They also added a pass to strip configuration info and add a utility function for querying MMA intrinsics, suggesting a focus on performance optimization and tuning for target hardware.
mlirspirvvulkantensorflowcompiler
bangtianliu/MatRox_RU
Aug 2019 - Dec 2019
Contributions:46 commits, 14 pushes 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.