Jacob Arndt
R&D Scientist at Oak Ridge National Laboratory
Oak Ridge, Tennessee, United States
Summary
🤩
Rockstar🎓
Top SchoolJacob Arndt is an R&D scientist with 12 years of experience applying geospatial data science, computer vision, and large-scale simulation techniques to earth observation and remote sensing problems. Based at Oak Ridge National Laboratory, he blends machine learning for high-resolution satellite imagery with deep expertise in high-performance scientific software—contributing to major open-source projects like the deal.II finite element library and Kokkos. His work spans end-to-end research and engineering: developing scalable land-use classification systems and optimizing parallel kernels and build systems for performance portability across CUDA, SYCL, and other backends. Jacob’s background in geography and computer science, plus rare experience in both dendrochronology fieldwork and performance-portable C++ development, lets him bridge domain science and low-level systems engineering effectively.
12 years of coding experience
9 years of employment as a software developer
Masters of Geographic Information Science Geographic Information Science, Masters of Geographic Information Science Geographic Information Science at University of Minnesota
Stackoverflow
11reputation
245reached
0answers
1question
Github Skills (25)
programming-language10
c-language10
package-management10
mpi10
cluster-computing10
cmake10
sycl10
parallel-computing10
performance-optimization10
accelerated-computing10
cuda10
build-tools10
high-performance10
cprogramming-language10
linear-algebra10
Programming languages (14)
Github contributions (5)
dealii/dealii
Sep 2013 - Jan 2023
The development repository for the deal.II finite element library
Role in this project:
Back-end Developer
Back-end DeveloperContributions:2202 reviews, 3865 commits, 3428 PRs in 9 years 6 months
Contributions summary:Jacob has contributed to the deal.II finite element library by modifying and optimizing the C++ code in the source files. Their commits focus on utilizing CUDA-aware MPI in the `Vector::compress*` functions, and in other cases improving the handling of the data in the implementation. Further improvements are made to the functions in order to deal with potentially empty ranges in CUDA kernels. These changes show a deep understanding of the libraries structure.
dealmultiphysicsc-plus-pluscomputational-sciencefinite-elements
kokkos/kokkos
Feb 2019 - Jan 2023
Kokkos C++ Performance Portability Programming Ecosystem: The Programming Model - Parallel Execution and Memory Abstraction
Role in this project:Back-end Developer
Back-end DeveloperContributions:4425 reviews, 1433 commits, 1165 PRs in 3 years 11 months
Contributions summary:Jacob primarily focused on improving the performance of the SYCL parallel_reduce functionality within the Kokkos project, implementing and testing new features. They introduced overloads, made performance optimizations by altering workgroup sizes and introduced a more general design, and addressed memory access problems. The user's work also included fixing bugs related to code behavior, particularly with edge cases for the team reduction calls.
memorympic-plus-plusmulti-threadingkokkos
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.