I published a short, interactive explainer that shows how a tiny neural network learns a linear separator. The page includes two side-by-side demos (supervised training and interactive feedback) that run entirely in your browser.

Read the post

Brief overview of what the 5.7 update of Unreal Engine means for the Synavis implementation.

Read the Unreal Engine update

For the next stage of the migration, I decided to use libdatachannel as the backend.

Brief post, read more here.

I published a short opinion piece about how I use AI tools in programming — when they help and when they don't. A concise read for anyone curious about practical AI-assisted coding.

Read the AI assistance post

Synavis has been using the (legacy) PixelStreaming1 plugin for a while now. This limits some aspects of the design space, particularly surrounding multiple data streams and on-demand rendering of additional synthetic streams. The migration is at a crossroads as there are quite technical considerations to be assessed.

Full article

Curious about how I create videos to accompany my research papers? Check out my latest blog post detailing my workflow, tools, and design choices for academic video production.

Read about video production

Our paper is now published in in silico Plants! The work details our efforts to make the virtual world rendering into an actual virtual world model, with tangible benefits to both the synthetic data as well as the simulation.

This paper is the culmunation of the efforts to react to common pitfalls of graphics-engine-based pipelines to generate synthetic data and it can be accessed here: 10.1093/insilicoplants/diaf018.

Welcome to my website! This site serves as a platform to share my research, publications, and updates in the field of computational engineering. The caviat is that this is not a CMS - and as such, older things will be removed and new things will be added as I go along.

Thank you for visiting, and I hope you enjoy exploring my work!

I have successfully defended my doctoral thesis at the University of Iceland! This marks the completion of several years of research in computational plant modeling and data visualization.

Thank you to my advisors, committee, and collaborators for their invaluable support.

Read more about the defense

I am excited to announce the publication of our latest research paper titled "VRoot: A VR-Based Application for Manual Root System Architecture Reconstruction" in Plant Phenomics. This work presents a novel virtual reality application designed to facilitate the manual reconstruction of root system architectures, enhancing the accuracy and efficiency of root phenotyping.

The paper is co-authored with my colleagues from Forschungszentrum Jülich and can be accessed here.