Sylwester Fraczek
Low Level Software Engineer at Nord Security
Gdańsk, Pomeranian Voivodeship, Poland
Summary
🤩
Rockstar🎓
Top School
Top expert inDeep Learning and Computer Vision Technologies
Sylwester Fraczek is a Low Level Software Engineer with a decade of experience specializing in deep learning framework development and hardware-accelerated backends. He has driven performance and integration work on major open-source projects like Apache MXNet and PaddlePaddle, focusing on oneDNN/MKL optimizations, quantization fixes, and operator-level improvements. At Intel he spent nearly a decade building and optimizing runtimes and operators for frameworks and accelerators (including Habana Gaudi), blending C++ and Python in Linux environments. Now based in Gdańsk and currently at Nord Security, he brings systems-level rigor to production ML stacks and build tooling. Colleagues value his ability to translate low-level changes into measurable runtime gains, and his background in Intelligent Decision Systems plus automation and robotics gives him an uncommon edge in algorithmic thinking for infrastructure work.
10 years of coding experience
10 years of employment as a software developer
Master’s Degree Intelligent Decision Systems, Master’s Degree Intelligent Decision Systems at Gdańsk University of Technology
Polish, English, Japanese
Github Skills (16)
paddlepaddle10
quantization10
mxnet10
machine-learning10
c-language10
deep-learning10
cprogramming-language10
python10
edn9
back-end-development9
neural-network9
convolutional-neural-networks9
compiler8
testing8
cmake7
Programming languages (8)
Github contributions (5)
PaddlePaddle/Paddle
Jul 2018 - Nov 2022
PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:187 reviews, 72 commits, 76 PRs in 4 years 5 months
Contributions summary:Sylwester contributed to the PaddlePaddle deep learning framework by focusing on improvements related to the MKLDNN (oneDNN) backend. Their work involved optimizing existing operations like top-k, convolution, and elementwise operations, and adding support for new features, such as the Swish activation function, within the MKLDNN framework. Furthermore, they were involved in testing and ensuring the correct behavior of these operations, including integration with the unit test suite, and in one case contributed to the code relating to the quantization procedure of the framework.
pytorchpythonparalleldeep-learningpaddlepaddle
apache/mxnet
Sep 2020 - Oct 2021
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Role in this project:ML Engineer
ML EngineerContributions:31 reviews, 8 commits, 8 PRs in 1 year
Contributions summary:Sylwester primarily contributed to the maintenance and improvement of the MXNet deep learning framework. Their commits included bug fixes related to quantization, particularly around the `concatenate` and `fc_eltwise` operations, and adjustments to the oneDNN integration. They also updated build scripts and documentation related to MKLROOT settings and the oneDNN integration. This suggests a focus on optimizing the framework's performance and integration with hardware acceleration libraries.
pythonschedulerdataflowmutationdata-science
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.