Peng Tu
System Architect Of Deep Learning, Computer Vision And Automotive Software Toolchain at Intel Corporation
San Ramon, California, United States
Summary
🤩
Rockstar🎓
Top SchoolPeng Tu is a seasoned system architect and compiler expert with 13+ years at Intel designing deep learning, computer vision, and automotive toolchains. He brings deep technical mastery in compiler backends—loop optimizations, vectorization, parallelization and scalar optimizers—rooted in a Ph.D. in computer science from UIUC and decades of optimizer/codegen work dating back to SGI and Tensilica. At Intel he has led cross-platform developer experiences (INDE) and continues to architect production toolchains that bridge research-grade optimizations with automotive and ML deployment needs. An active contributor to the widely used ISPC project, he has directly fixed complex codegen issues around lane management and LLVM integration, demonstrating pragmatic expertise with SPMD and LLVM internals. Based in the Bay Area, he pairs deep compiler internals knowledge with product-minded engineering to deliver optimizations that matter in performance-critical systems.
13 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Computer Science, Doctor of Philosophy (Ph.D.), Computer Science at University of Illinois Urbana-ChampaignMaster, Computer Science, Master, Computer Science at Shanghai Jiao Tong University
Github Skills (9)
compiler-design10
c-language10
cprogramming-language10
llvm10
optimization9
programming-language9
simd9
spmd8
intel8
Programming languages (7)
Github contributions (5)
ispc/ispc
Oct 2012 - Feb 2013
Intel® Implicit SPMD Program Compiler
Role in this project:
Backend Developer
Backend DeveloperContributions:8 commits in 4 months
Contributions summary:Peng primarily focused on modifying the ISPC compiler's core code generation, specifically around loop unrolling and lane management. They addressed issues in the `stmt.cpp` file related to the correct placement of lane restoration within foreach loops and updated internal LLVM usage within the compiler. Additional contributions included adjustments to optimization passes and fixing build issues related to LLVM integration, demonstrating a strong understanding of compiler internals.
cpusimdavximplicitavx512
pengtu/triton
Mar 2023 - Aug 2023
Development repository for the Triton language and compiler
Contributions:2 PRs, 27 pushes, 4 branches in 5 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.