I am a self-taught C++ developer. I am currently an undergraduate student at the University of Warsaw. My main area of expertise is high performance work with GPUs, including accelerating machine learning and realtime graphics rendering.
Recently, I have interned twice at NVIDIA, working on NVIDIA Transformer Engine – the library for accelerating LLM training using newest architectural features. I also gave a talk on performance at The International C++ Conference. A distributed systems paper I co-authored has been published at OSDI. You can find some of my projects on Github. I also host an educational YouTube channel, where I share my knowledge with the world. There are also some informative posts on this website.
Previously, in high school, I took part in Cyberduck – a robotics project whose goal was to create a portable aqueous meteorological station in the shape of a mallard. Additionally, I also participated in PaSh – a shell-script parallelisation project hosted by the Linux Foundation.
In free time, I like doing renders in Blender – an open-source 3D graphics suite. You can find my works in the Gallery. At the same time, you can see some images created with renderers of my own making at the bottom of the page.
Jan Bielak’s Publications
Practically Correct, Just-in-Time Shell Script Parallelization
Konstantinos Kallas, Tammam Mustafa, Jan Bielak, Dimitris Karnikis, Thurston H.Y. Dang, Michael Greenberg, Nikos Vasilakis. 16th USENIX Symposium on Operating Systems Design and Implementation (OSDI22)
Media Coverage
- Faster computing results without fear of errors – MIT News’ coverage of PaSh, June 7th 2022
Talks
- Your Performance Todo List – CppCon 2022, September 14th 2022 (view on CppCon YouTube) (interactive slide deck)
Jan Bielak’s Rendering Examples



