Petr Kurapov
Software Engineer at Intel Corporation
Munich, Bavaria, Germany
Summary
🤩
Rockstar🎓
Top SchoolPetr Kurapov is a software engineer with 12 years' experience based in Munich and a PhD-level researcher from MIPT, currently at Intel working on high-performance systems and analytical query processing. He contributes to flagship open-source compiler projects like ISPC—implementing Level Zero kernel timing, increased parallelism and GPU dispatch fixes—and LLVM/MLIR, adding XeGPU dialect features and tensor sg_map support to enable work-item semantics. His background uniquely blends low-level systems engineering (kernel/driver work, PMU-based anomaly detection) with applied security research, including ML-enhanced side-channel analysis and automated oscilloscope data pipelines. That rare combination of compiler backend tuning, GPU acceleration, and security instrumentation lets him deliver measurable, real-time performance improvements across hardware and software stacks.
12 years of coding experience
8 years of employment as a software developer
Moscow Institute of Physics and Technology (State University) (MIPT)
English, Russian
Github Skills (24)
performance-monitor10
gpu-programming10
performanceanalysis10
mlr10
performance-measurement10
compiler-compiler10
performance-analysis10
gpu10
performanceanalytics10
performance-tuning10
compiler10
performance-monitoring10
ispc10
performancemonitor10
compilers10
Programming languages (9)
Github contributions (5)
llvm/llvm-project
Dec 2022 - Mar 2025
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:61 reviews, 27 PRs, 3 pushes in 2 years 3 months
Contributions summary:Petr's contributions center around adding and modifying features within the MLIR (Multi-Level Intermediate Representation) framework, specifically focusing on the XeGPU dialect. They implemented a `select` operation within the Linalg dialect and added the `sg_map` attribute for tensor descriptors to support work item-level semantics in the XeGPU dialect, improving the capabilities for hardware acceleration. Furthermore, the user made changes to accommodate the `sg_map` attribute in existing GPU-related operations like `load_nd` and `store_nd`. These contributions directly relate to enhancing the framework's functionality for GPU compilation and optimization.
compilerstechnologiesclangsubmittoolchain
ispc/ispc
Oct 2019 - Apr 2020
Intel® Implicit SPMD Program Compiler
Role in this project:Performance Engineer
Performance EngineerContributions:24 commits in 6 months
Contributions summary:Petr primarily focused on performance optimization within the ISPC compiler repository. Their contributions included integrating Level Zero (L0) API for kernel time measurement, adding more parallelism to the aobench example, and removing unnecessary synchronizations. They also updated the code to the latest L0 API and fixed dispatch issues in existing examples, suggesting a focus on efficient GPU utilization.
cpusimdavximplicitavx512
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.