Summary
Abhishek Yenpure is a Senior Systems Software Engineer and HPC/GPU specialist with a Ph.D. in Computer Science and nine years of experience optimizing CUDA kernels, distributed MPI systems, and large-scale scientific visualization. He has scaled visualization and analytics pipelines to over 8,000 GPUs on DOE supercomputers and delivered performance-portable algorithms across CUDA, HIP, and SYCL backends via Kokkos. At Kitware he combined C++ template metaprogramming with low-level profiling (nvprof, VTune, PAPI) to squeeze memory efficiency and throughput from VTK-m and ParaView, and now contributes to GPU communications at NVIDIA (NCCL/NVSHMEM) for large-scale distributed GPU computing. His work spans end-to-end HPC stacks—from kernel tuning and hybrid CPU–GPU execution to in situ analytics and domain-specific Python/xarray tooling for climate, fusion, and geospatial science. He is an active open-source contributor who pairs deep research experience with production engineering, and his background in building Exascale-ready pipelines is often the hidden accelerator behind collaborative scientific breakthroughs.
9 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy - PhD Scientific Visualization and High Performance Computing, Doctor of Philosophy - PhD Scientific Visualization and High Performance Computing at University of Oregon
Bachelor of Engineering (B.E.) Information Technology, Bachelor of Engineering (B.E.) Information Technology at Savitribai Phule Pune University
Marathi, Hindi, English