Anna Karbownik
HPC Software Engineer at ParTec AG
Feldkirchen-Westerham, Bavaria, Germany
Summary
🤩
Rockstar🎓
Top SchoolAnna Karbownik is an HPC Software Engineer with eight years of experience building and optimizing high-performance computing and deep learning stacks, currently based in Feldkirchen-Westerham, Germany. Her background at Intel spans roles from graphics and driver development to deep learning framework work and AI application enablement, giving her a rare blend of low-level systems expertise and ML deployment know-how. She is an active contributor to the Apache MXNet project, where she tackled MKLDNN BatchNorm bugs and improved cross-platform build and MKL integration, reflecting strong DevOps and cross-platform build skills. At ParTec and Intel she has focused on performance tuning, drivers and platform enablement for CPUs and accelerators, often bridging R&D and customer-facing engineering. Colleagues rely on her for pragmatic debugging of complex performance issues and for unglamorous but critical build and integration work that keeps production ML systems running.
8 years of coding experience
16 years of employment as a software developer
Master's degree Telecommunications Engineering, Master's degree Telecommunications Engineering at Gdańsk University of Technology
Github Skills (10)
mxnet10
build-system10
cmake10
python9
machine-learning9
deep-learning9
openmp9
c-language8
cicd8
cprogramming-language8
Programming languages (2)
Github contributions (5)
apache/mxnet
Oct 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:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:42 reviews, 10 commits, 49 PRs in 1 year
Contributions summary:Anna contributed to bug fixes, particularly addressing issues in the MKLDNN BatchNorm implementation related to channel numbers. They also worked on improving the build process and supporting Ubuntu 20.04, including integrating Intel MKL libraries. Furthermore, they modified the build system, removing an OpenMP submodule and enabling dynamic linking with MKL and compiler-based OpenMP, particularly for MacOS. The user also addressed a test failure and refactored code.
pythonschedulerdataflowmutationdata-science
akarbown/incubator-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
Contributions:80 pushes, 20 branches, 1 tag in 1 year 1 month
pythonschedulerdataflowmutationorchestration
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.