Konrad Zawora
Senior AI Software Engineer at (can’t say)
Gdańsk, Pomeranian Voivodeship, Poland
Summary
🤩
Rockstar🎓
Top SchoolKonrad Zawora is a Senior AI Software Engineer based in Gdańsk with a strong track record of low-level AI software and performance engineering for next-generation accelerators. Over ~3 years of professional experience and prior internships, he led hands-on efforts at Intel and Habana Labs to bring Intel Gaudi support to vLLM, optimizing inference backends and HPU-specific graphs for high-throughput LLM serving. He combines deep systems expertise—compiler, math kernels, and pre-silicon GPU work—with practical MLOps know-how in configuring builds and fixing multiprocessing and device-specific issues. As a research and teaching assistant at Gdańsk University of Technology, he continues to bridge academia and industry while exploring new technical directions. Notably, his open-source contributions to the well-known vLLM project emphasize hardware-aware optimizations that accelerate real-world model inference. Currently focused on an exciting new chapter and not seeking opportunities, he prefers building high-performance, hardware-integrated AI software.
3 years of coding experience
4 years of employment as a software developer
Master of Engineering - MEng Computer Science, Master of Engineering - MEng Computer Science at Gdańsk University of Technology
English, Polish, German
Github Skills (12)
pytorch10
mlops10
inference10
dockers9
docker9
develop8
cuda8
backend8
transformer7
roc7
amd6
deepspeed5
Programming languages (1)
Github contributions (5)
vllm-project/vllm
Dec 2023 - Mar 2025
A high-throughput and memory-efficient inference and serving engine for LLMs
Role in this project:
MLOps Engineer
MLOps EngineerContributions:41 reviews, 8 PRs, 45 comments in 1 year 2 months
Contributions summary:Konrad's contributions primarily focus on integrating and optimizing the vLLM project for Intel Gaudi (HPU) hardware, specifically addressing inference backends. Their work involves modifying the codebase to support HPU-specific configurations, including the implementation of HPU graphs for performance acceleration. The user is also responsible for configuring build environments and fixing issues related to HPU multiprocessing and the block size for Gaudi devices, showing a deep understanding of the project's infrastructure and hardware integration.
amdcudadeepseekgpthpu
HabanaAI/vllm-fork
Dec 2023 - Apr 2025
A high-throughput and memory-efficient inference and serving engine for LLMs
Contributions:3 releases, 185 reviews, 335 PRs in 1 year 3 months
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.