Karthik Jayaprakash
Senior Member Of Technical Staff at AMD
Boxborough, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolKarthik Jayaprakash is a Senior Member of Technical Staff at AMD with six years of focused experience in heterogeneous compute software and runtime systems. He develops and architects backend components for ROCm/HIP and the Compute Language Runtime, contributing critical fixes and features such as hipFuncGetAttribute, improved module handling, and clearer error reporting. His background spans low-level modem and PHY/Layer work at Qualcomm to stream compute at AMD, giving him a strong systems-and-performance-first engineering perspective. Based in Boxborough, MA, he blends production-grade implementation skills with platform design, frequently addressing cross-layer compilation and module-loading challenges. An active open-source contributor to the widely used ROCm/hip project, he helps bridge hardware capabilities and developer-facing APIs. Colleagues rely on him for pragmatic solutions that improve portability and debuggability across heterogeneous compute stacks.
6 years of coding experience
14 years of employment as a software developer
Masters of Science Computer Science, Masters of Science Computer Science at Binghamton UniversityBachelors of Engineering Computer Science and Engineering, Bachelors of Engineering Computer Science and Engineering at Velammal Engineering College
Tamil
Github Skills (11)
module-loader10
error-handling10
c-language10
lazy-loading10
cprogramming-language10
resource-loading10
architecture9
architectures9
system9
clang8
clangd8
Programming languages (2)
Github contributions (5)
ROCm/hip
Oct 2019 - Aug 2022
HIP: C++ Heterogeneous-Compute Interface for Portability
Role in this project:
Back-end Developer & System Architect
Back-end Developer & System ArchitectContributions:3 reviews, 67 commits, 10 PRs in 2 years 10 months
Contributions summary:Karthik primarily contributed to the ROCm/HIP project by fixing compilation errors, enhancing the underlying platform's functionality, and addressing issues related to texture references and module loading. They introduced changes to the internal platform state, module handling, and error code management within the HIP layer. Additionally, the user implemented and integrated support for new features, such as the `hipFuncGetAttribute` and improved the quality of error messages in the HIP layer.
cudaheterogeneousgpuportabilityhip-runtime
kjayapra-amd/HIP
Sep 2019 - May 2022
HIP : Convert CUDA to Portable C++ Code
Contributions:6 pushes, 18 branches in 2 years 8 months
cudacppgpuc-plus-plusportable
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.