Artem Polyakov
Staff Architect at NVIDIA
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolArtem Polyakov is a Staff Architect with 19 years of deep expertise in Linux kernel and system software, high-performance computing (HPC) stacks, and HPC interconnects, currently driving architecture at NVIDIA in New York. His career spans research and industry roles at Mellanox and ISP SB RAS, where he progressed from kernel and system-level development to leading performance-sensitive networking and software architecture efforts. Artem is a hands-on contributor to major open-source HPC projects—Open MPI and UCX—delivering bug fixes, race-condition hardening, and low-latency optimizations for InfiniBand and RC transports. He also contributed visualization and tooling improvements to DMTCP’s inspector, showing a pragmatic focus on both developer ergonomics and production performance. Holding a PhD in Computer Systems, he combines rigorous academic training with a track record of shipping low-level, hardware-aware optimizations that materially improve cluster communication latency and reliability.
19 years of coding experience
15 years of employment as a software developer
Doctor of Philosophy (PhD), Computer Systems, Doctor of Philosophy (PhD), Computer Systems at Siberian State University of Telecommunications and Information Sciences
Russian, English
Github Skills (24)
infiniband10
performance-analytics10
performance-monitor10
checkpoint10
c-language10
rdma10
mpi10
c1110
networking10
performance-measurement10
checkpointing10
c1710
performance-analysis10
rest10
graph-visualization10
Programming languages (6)
Github contributions (5)
dmtcp/dmtcp
Jul 2008 - Mar 2016
DMTCP: Distributed MultiThreaded CheckPointing
Role in this project:
Back-end Developer
Back-end DeveloperContributions:227 commits, 4 PRs, 12 pushes in 7 years 9 months
Contributions summary:Artem primarily contributed to the `dmtcpinspector` tool, focusing on enhancements to visualize the relationships between processes within the DMTCP framework. Their work involved refactoring the code to group host processes in subgraphs, modifying the visual representation of connections, and adding features like parent-child relationship display. They also implemented improvements to the build process and added new command line options for controlling the display.
mpiinfinibanddmtcpcheckpoint-restartcluster-computing
open-mpi/ompi
Sep 2014 - Sep 2021
Open MPI main development repository
Role in this project:Back-end Developer
Back-end DeveloperContributions:15 reviews, 219 commits, 164 PRs in 7 years
Contributions summary:Artem primarily focused on improving the Open MPI's functionality through various bug fixes and enhancements. The contributions involved correcting issues in intercommunicator splitting logic, solving problems related to MPI_Alltoall for inter-communicators, iBarrier for inter-communicators and iGatherv. The user also addressed memory registration limits and race conditions within the OpenIB BTL (Byte Transfer Layer), improving performance and reliability in network communication. Furthermore, the user worked on code cleanup and integration with UCX for the same library.
mpicluster-computingfortranopenmpipetsc
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.