Allard Hendriksen
Senior Developer Technology Engineer at NVIDIA
Munich, Bavaria, Germany
Summary
🤩
Rockstar🎓
Top SchoolAllard Hendriksen is a Senior Developer Technology Engineer based in Munich with a decade of experience applying advanced numerical methods and machine learning to real-world systems. Holding an MSc and PhD in Applied Mathematics from Leiden, he transitioned academic expertise in computational imaging into production-grade GPU-accelerated solutions at NVIDIA. He contributes to high-performance open-source projects like RAPIDS (RAFT), where his CUDA kernel optimizations and algorithmic improvements yield measurable speedups across GPU architectures. Comfortable bridging theory and engineering, he has a track record of squeezing performance from low-level code while keeping algorithms robust for deployment. Colleagues describe him as a detail-oriented problem solver who brings academic rigor to practical system constraints.
10 years of coding experience
4 years of employment as a software developer
MSc Applied Mathematics, MSc Applied Mathematics at Leiden UniversityCalifornia Institute of TechnologyGymnasium - Natuur en Techniek & Natuur en Gezondheid, Gymnasium - Natuur en Techniek & Natuur en Gezondheid at Christelijk Gymnasium Sorghvliet
Github Skills (16)
cuda10
machine-learning10
c-language10
cprogramming-language10
gpu10
performance-optimization10
atomic10
atomics10
linear-algebra9
data-structure9
algorithm9
data-structures9
algorithms9
sparse8
nearest-neighbors8
Programming languages (9)
Github contributions (5)
rapidsai/raft
Jul 2022 - Jan 2023
RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing high performance applications.
Role in this project:
ML Engineer
ML EngineerContributions:135 reviews, 6 commits, 33 PRs in 6 months
Contributions summary:Allard primarily focuses on optimizing and improving CUDA-accelerated algorithms within the RAPIDS AI ecosystem. Their contributions include implementing warp-aggregated atomic increments, which enhance performance in filtering operations. Additionally, they've replaced existing functions with faster equivalents and added launch bounds to kernel functions, specifically targeting the `adj_to_csr_kernel` function, indicating a focus on performance and compatibility in GPU computing contexts. They also addressed the performance of fusedL2NN when data is skinny, resulting in a performance improvement across GPU architectures.
sciencedata-sciencemachine-learninggraph-data-sciencefundamental
ahendriksen/tomosipo
Sep 2018 - Nov 2021
Contributions:270 commits, 4 PRs, 143 pushes in 3 years 1 month
cudareconstructionscientific-computingpython3tomography
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.