Janne Kiviluoto
Senior System Software Engineer at NVIDIA
Jyväskylä, Central Finland
Summary
🤩
RockstarJanne Kiviluoto is a Senior System Software Engineer with 16 years of experience, currently building system-level software at NVIDIA from his base in Jyväskylä, Finland. He brings deep low-level expertise gained across embedded, web, and systems roles dating back to early positions at Scandinavia Online, Nomovok, and Bluegiga. At NVIDIA he focuses on production-grade performance, reliability, and tooling, and has applied that mindset to open-source ML engineering work—contributing optimizations, TF32 support, CUDA kernel fixes, and improved error handling to the widely used StyleGAN2-ADA PyTorch repository. Janne combines pragmatic engineering with strong documentation and diagnostics skills, often improving developer experience as much as raw throughput. Known for quietly tackling thorny data-loading and metric-calculation bugs, he tends to improve systems from the inside out rather than through headline features. This blend of system-level rigor and hands-on ML tooling work makes him a valuable bridge between research code and production deployments.
16 years of coding experience
8 years of employment as a software developer
Github Skills (7)
cuda10
computer-vision10
pytorch10
machine-learning10
python10
docker8
dockers8
Programming languages (5)
Github contributions (5)
NVlabs/stylegan2-ada-pytorch
Feb 2021 - Feb 2021
StyleGAN2-ADA - Official PyTorch implementation
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 1 comment in 17 days
Contributions summary:Janne contributed to the optimization and enhancement of the StyleGAN2-ADA PyTorch implementation. Their work involved modifying the training pipeline, adding features to enable TF32 usage, and improving the error messages. The user also addressed issues related to data loading and improved documentation and made changes to how the metrics are calculated. They also addressed issues with the custom CUDA kernels, and enhanced the error handling.
pytorchadastylegancycleganimage-synthesis
NVlabs/noise2noise
Oct 2018 - Nov 2021
Noise2Noise: Learning Image Restoration without Clean Data - Official TensorFlow implementation of the ICML 2018 paper
Contributions:5 commits, 1 push in 3 years 1 month
restorationdeep-learningnoise2noiseimage-restorationclean
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.