Alexander Mikheev
Sr Staff Engineer at NVIDIA
Israel
Summary
🤩
Rockstar🎓
Top SchoolAlexander Mikheev is a Sr. Staff Engineer with 13+ years in high-performance computing and networking, now at NVIDIA after a long tenure at Mellanox where he worked on InfiniBand, RDMA, RoCE and storage/NVMe stacks. He architects and implements low-latency, high-throughput systems—contributing significant performance and reliability improvements to flagship open-source projects like UCX and Open MPI, including receiver-queue refactors, hardware tag matching, congestion avoidance, and SRQ/RNR optimizations. His background spans deep systems work from kernel drivers and network boot to distributed memory management and Ceph integration, reflecting a rare full-stack of HPC networking expertise. Based in Israel, he combines pragmatic engineering with system-level research instincts and a history of shipping production-grade features that tighten latency and resilience in cluster-scale deployments. Notably, his open-source contributions include both protocol-level innovations and extensive test coverage, underscoring a focus on robustness as well as performance.
13 years of coding experience
26 years of employment as a software developer
Bachelor’s Degree, Mathematics and Computer Science, Bachelor’s Degree, Mathematics and Computer Science at University of Haifa
English, Russian, Hebrew, Ukrainian, Turkish, Italian
Github Skills (27)
c-language10
parallel10
parallelization10
memory-management10
mpi10
hpc10
openmpi10
c1110
networking10
c1710
parallel-processing10
parallel-computing10
performance-optimization10
network-protocol10
protocols10
Programming languages (2)
Github contributions (5)
openucx/ucx
Nov 2014 - Feb 2018
Unified Communication X (mailing list - https://elist.ornl.gov/mailman/listinfo/ucx-group)
Role in this project:
Back-end Developer & System Architect
Back-end Developer & System ArchitectContributions:306 commits, 118 PRs, 303 comments in 3 years 3 months
Contributions summary:Alexander primarily contributed to improvements in the Unified Communication X (UCX) library related to InfiniBand (IB) and User Datagram Protocol (UD) transport layers. Their work included refactoring and optimizations of the receiver queue, adding support for hardware tag matching and the creation of endpoints within the UD context. Furthermore, the user implemented features related to dynamic connection and congestion avoidance algorithms, enhancing the robustness and performance of the networking library. The user's contributions also included implementing several tests to improve the reliability and verification of the networking layer.
craympiopenshmemyangroce
open-mpi/ompi
Jan 2013 - Mar 2018
Open MPI main development repository
Role in this project:Back-end Developer
Back-end DeveloperContributions:183 commits, 66 PRs, 2 pushes in 5 years 2 months
Contributions summary:Alexander's contributions primarily revolve around optimizing the Open MPI library, specifically within the openib BTL (BTL stands for Byte Transfer Layer) and related components. They fixed compilation warnings and optimized the handling of SRQ (Shared Receive Queue) settings to minimize RNRs (Receiver Not Ready errors). Furthermore, the user made modifications to memory management, including improvements to memory registration and the handling of memory keys within the OSHMEM (OpenSHMEM) component, addressing issues related to memory allocation, and inter-process communication.
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.