Konrad Dobroś
Senior Software Developer at Luxoft
Gdańsk, Pomeranian Voivodeship, Poland
Summary
🤩
Rockstar🎓
Top SchoolKonrad Dobroś is a senior software engineer with six years of experience specializing in back-end development and performance engineering for ML and inference platforms. He has contributed significant optimizations to OpenVINO’s clDNN kernels and extended PlaidML to model device-host interactions, demonstrating deep hands-on expertise with OpenCL, SPIR-V and GPU memory/synchronization. Konrad has held roles at Intel and Luxoft before joining AMD as a Senior Member of Technical Staff, combining production-focused engineering with low-level performance tuning. Based in Gdańsk, he brings a practical knack for reducing device allocations and improving kernel performance—skills that sit at the intersection of compiler-level transformations and runtime efficiency.
6 years of coding experience
6 years of employment as a software developer
Inżynier (Inż.) Informatyka, Inżynier (Inż.) Informatyka at AGH University of Krakow
English, Polish
Github Skills (19)
spirv10
opencl10
c-language10
gpu-programming10
inference10
intermediate10
multilevel10
deep-learning10
performance-optimization10
cprogramming-language10
deep-q-learning9
optimizer9
ai9
optimisation9
optimization9
Programming languages (2)
Github contributions (5)
plaidml/plaidml
Sep 2020 - Nov 2020
PlaidML is a framework for making deep learning work everywhere.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:21 reviews, 22 commits, 24 PRs in 2 months
Contributions summary:Konrad contributed to the PlaidML project, a framework for deep learning. Their work focused on extending the "comp" dialect to model device-host interactions, including memory management and synchronization. They also added OpenCL third-party archives, and implemented OpenCL runtime and runner, demonstrating experience with OpenCL and SPIR-V. Additionally, they enhanced the compilation process by adding code to convert GPU to comp, and contributed to reducing device allocation operations.
pytorchtvmdeep-learningmachine-learningcompiler
openvinotoolkit/openvino
May 2020 - Sep 2020
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference
Role in this project:Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:13 reviews, 22 commits, 21 PRs in 3 months
Contributions summary:Konrad primarily contributed to optimizing the clDNN (Compute Library for Deep Neural Networks) component of the OpenVINO toolkit. Their work involved enhancing the performance of specific kernels, such as 1x1 imad convolution, and addressing issues related to data format conversions (fsv16). They also added support for asymmetric quantization and improved the handling of data in various formats. Furthermore, the user implemented optimizations for the Gather primitive and corrected dimension calculations.
inference-enginepytorchmodel-optimizerdeep-learninggpu
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.