Joe Todd
Tampa, Florida United States
Summary
🤩
Rockstar🎓
Top SchoolJoe Todd is a senior consultative technologist with 20+ years blending IT, insurance finance and research administration expertise into practical, cross-industry solutions. He is an experienced C++ back-end developer and performance engineer who has spent recent years porting and optimizing GPU code with SYCL and CUDA—contributing notable improvements to high-profile open-source projects like LAMMPS and Kokkos. Joe’s work focuses on performance, memory management and stability—implementing SYCL team reductions, USM memory pools, and pinned host memory integrations to enable robust GPU-accelerated simulations. Based in Tampa, he pairs technical depth with strong relationship-building and client-facing skills, often translating complex computational problems into deployable results. A former author of 3D iceberg calving models, he brings a rare mix of scientific modeling insight and systems-level performance tuning to engineering challenges.
11 years of coding experience
Bachelor of Science (BS), Accounting and Computer Science, Bachelor of Science (BS), Accounting and Computer Science at Southern New Hampshire UniversityFox Lane High School - Bedford NY
Github Skills (15)
sycl10
gpu-programming10
cluster-computing10
programming-language10
c-language10
molecular-dynamics-simulation10
parallel-computing10
cprogramming-language10
molecular-simulation10
scientific-computing10
k10
dynamics10
performance-optimization9
cmake9
openmp8
Programming languages (9)
Github contributions (5)
kokkos/kokkos
Oct 2021 - Feb 2022
Kokkos C++ Performance Portability Programming Ecosystem: The Programming Model - Parallel Execution and Memory Abstraction
Role in this project:
Back-end Developer
Back-end DeveloperContributions:15 reviews, 27 commits, 5 PRs in 3 months
Contributions summary:Joe focused on improving the SYCL backend within the Kokkos framework. Their contributions primarily involved addressing lifetime issues, optimizing performance, and implementing a USM (Unified Shared Memory) pool for memory allocations. They made several improvements to the SYCL integration, including adjustments to memory management, copy operations, and the handling of kernel events. These changes reflect a focus on enhancing performance and stability in the SYCL implementation.
cppparallel-executionprogramming-modelparallel-computingc-plus-plus
lammps/lammps
Jul 2021 - Mar 2022
Public development project of the LAMMPS MD software package
Role in this project:Back-end Developer / Performance Engineer
Back-end Developer / Performance EngineerContributions:5 commits, 4 PRs, 6 comments in 8 months
Contributions summary:Joe's contributions primarily involve adapting the LAMMPS molecular dynamics software to utilize SYCL for GPU acceleration. They implemented SYCL team_reduce to mimic CUDA's __syncthreads_count and fixed related potential out-of-bounds errors within the GPU kernels. Furthermore, the user integrated SYCL pinned host memory from Kokkos and added cmake configuration files for SYCL builds. These changes suggest a focus on improving performance and supporting SYCL-based computations within the LAMMPS simulation environment.
lammpsmolecular-dynamicssimulationkokkos
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.