Mads Kristensen
Principal Software Engineer at NVIDIA
Copenhagen, Capital Region of Denmark
Summary
🤩
Rockstar🎓
Top SchoolMads Kristensen is a Principal Software Engineer based in Copenhagen with 15 years of experience building high-performance, distributed systems and GPU-accelerated data tooling. Now at NVIDIA, he brings deep expertise in parallel computing and memory management, having made notable open-source contributions to Dask, cuDF, RMM and Numba that improved scheduling, serialization, and GPU memory allocation. His background spans academia to industry—PhD-trained and formerly an assistant professor—giving him a strong foundation in research-driven engineering and rigorous testing. Colleagues rely on him for thoughtful refactors and edge-case fixes that stabilize core libraries used in large-scale data processing, and he often bridges CUDA-centric performance work with practical developer ergonomics. An understated strength is his track record of integrating cutting-edge GPU allocators into mainstream Python ecosystems, enabling real-world speedups for data-intensive workloads.
15 years of coding experience
11 years of employment as a software developer
Ph.D, Computer Science, Ph.D, Computer Science at Københavns Universitet - University of Copenhagen
Danish, English
Github Skills (31)
memory-allocation10
msgpack10
python10
testing10
memory-management10
cudf10
data-serialization10
numpy10
distributed-computing10
serialization10
rapids10
compiler-compiler10
dask10
cupy10
cuda10
Programming languages (8)
Github contributions (5)
dask/dask
May 2019 - Oct 2024
Parallel computing with task scheduling
Role in this project:
Back-end Developer
Back-end DeveloperContributions:80 reviews, 60 PRs, 177 comments in 5 years 5 months
Contributions summary:Mads made several contributions to the Dask library, focusing on optimizing and maintaining the core codebase. Their work includes removing and updating dependencies, specifically relating to the `cudf` library. They also implemented fixes for HighLevelGraph dependencies and performed code refactoring by moving functions and updating API calls, which improved the efficiency and stability of the library. In addition, the user addressed a few edge cases related to the `set_index` operation and serialization of the `Blockwise` layer.
pythonschedulingparallelnumpydask
dask/distributed
Oct 2020 - Jan 2023
A distributed task scheduler for Dask
Role in this project:Back-end Developer
Back-end DeveloperContributions:79 reviews, 26 commits, 62 PRs in 2 years 2 months
Contributions summary:Mads contributed to the `dask/distributed` repository, a distributed task scheduler for Dask. The commits indicate a focus on enhancing the core functionality of the scheduler, specifically by implementing support for HighLevelGraphs, which are crucial for task scheduling. The user also made changes to improve the serialization and communication protocols. The user's work directly impacted how tasks are processed and managed within the distributed system.
pythonschedulerdaskdistributed-computingdistributed-systems
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.