Michał Krzywkowski
Senior Software Engineer at NVIDIA
Warsaw, Masovian Voivodeship, Poland
Summary
🤩
RockstarMichał Krzywkowski is a senior software engineer based in Warsaw with 10 years of experience building high-performance, networked systems and low-level tooling. He specializes in C, C++ and Python on Linux, with deep practical experience in packet processing, DDoS mitigation, traffic analysis and performance-sensitive financial data pipelines. At NVIDIA he contributes to developer tooling for CUDA and has back-end/MLOps experience improving the GPU-accelerated DALI library—work that included optimizing video decoding, expanding operators, and fixing memory/synchronization issues. His background spans embedded and OS-level projects (Tizen memory analysis) through to production-scale networking at EXATEL, showing a consistent focus on reliability, efficiency and security. Colleagues benefit from his knack for finding subtle race conditions and squeezing extra performance out of constrained systems.
10 years of coding experience
7 years of employment as a software developer
English, Polish
Github Skills (15)
tensorrt10
cuda10
tensor10
c-language10
parallel-computing10
cprogramming-language10
operation10
image-processing10
testing9
tensorflow9
performance-optimization9
build-system9
cmake9
machine-learning8
ml8
Programming languages (9)
Github contributions (5)
NVIDIA/DALI
Nov 2018 - Jan 2023
A GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.
Role in this project:
Back-end Developer & MLOps Engineer
Back-end Developer & MLOps EngineerContributions:5881 reviews, 439 commits, 1237 PRs in 4 years 2 months
Contributions summary:Michal Zientkiewicz focused on enhancing the DALI library, contributing to its core functionality through various improvements. His work included optimizing the performance and handling of video decoding by addressing synchronization issues, and expanding the operators available. He also implemented support for image indexing and added a range of features to several existing operations. He also made some changes to tests to increase coverage and address found memory issues.
deep-learninggpuinferencefast-data-pipelineimage-augmentation
mzient/genFFT
Jul 2017 - May 2021
Template-based FFT, optimized for x86 processors.
Contributions:28 commits, 2 PRs, 20 pushes in 3 years 11 months
fftx86
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.