Todd Anderson
Principal Software Engineer at bodo.ai
Hillsboro, Oregon, United States
Summary
🤩
Rockstar🎓
Top SchoolTodd Anderson is a Principal Software Engineer and researcher with roughly 17 years focused on programming systems, parallelism, and scaling computations across CPUs, GPUs, and clusters. He created the auto-parallelization feature in Numba—used by hundreds of thousands and responsible for major speedups—and authored PyOMP and Ramba to bring OpenMP semantics and distributed NumPy workloads to Python with near-C performance. His recent work applies those auto-distribution techniques to tensor placement for LLM training and uses AI for program synthesis, implementation selection, and performance prediction across thousands of candidate kernels. Based in Hillsboro, Oregon, Todd blends deep compiler and runtime expertise (including managed runtimes and GC) with practical open-source impact, evidenced by core contributions to numba/numba. He routinely turns research ideas into developer-facing tools that significantly reduce programmer effort while improving cost and performance.
15 years of coding experience
26 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Computer Science, Doctor of Philosophy (Ph.D.), Computer Science at University of KentuckyBachelor of Science (B.S.), Computer Science, Bachelor of Science (B.S.), Computer Science at Marshall University
Github Skills (13)
list-processing10
compiler10
numba10
array-map10
parallel-computing10
jit10
python10
compile10
numpy10
llvm10
code-optimization9
cprogramming-language7
c-language7
Programming languages (10)
Github contributions (5)
numba/numba
Feb 2017 - Jan 2023
NumPy aware dynamic Python compiler using LLVM
Role in this project:
Back-end Developer
Back-end DeveloperContributions:193 reviews, 705 commits, 131 PRs in 5 years 11 months
Contributions summary:Todd made contributions to improve the functionality of Numba, a NumPy aware dynamic Python compiler using LLVM. The contributions focused on enhancing the compilation process by adding support for new language features like array slicing, optimizing loop behavior by implementing a memory model, and streamlining array reductions in a parallel fashion. The work involved extending the compilation pipeline, including adding code to manage and enhance the performance of the core compilation operations.
cudapythonparallelnumpynumba
Python-for-HPC/numbaWithOpenmp
Mar 2022 - Nov 2024
Contributions:49 reviews, 30 PRs, 155 pushes in 2 years 9 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.