Recurrent Flow Networks
Proposing an unsupervised algorithm to generalise the underlying distribution of video data.
- Python
- PyTorch
Portfolio
About
I live in Copenhagen, where I create autonomous systems.
As a kid I was fascinated by space, nature, and science. I spent hours tinkering with computers, observing ants, and reading about technology and mathematics.
I've always been captivated by space - a vast, dark expanse where you can gaze at the stars and let your thoughts drift. It always invites existential questions; What's out there? Why is it there? Why are we here? I credit that fascination with the unknown for my love of science fiction, science, and even the design of this website.
In my teenage years, I took up strength training. I wanted to learn anatomy and physiology - and like most teenagers, I had ambitions involving bigger biceps. That journey led me into powerlifting and eventually to becoming a certified strength coach. The same drive to understand how things work - whether muscles and tendons or neurons in a neural network - has shaped my path ever since. It ultimately led me to pursue a career at the intersection of research and engineering, where I later specialized in generative machine learning.
Today, I'm a Senior AI Specialist, building AI systems and experimenting with emerging technologies. In my free time, I code, hike, lift heavy things, and spend time with my family and friends.
I love learning and connecting with curious people. If you'd like to chat about a project, build something together, or just exchange ideas, feel free to reach out!
Projects
Proposing an unsupervised algorithm to generalise the underlying distribution of video data.
Visualising and analysing the relationship between COVID-19 and socioeconomic status.
A modern, responsive portfolio website built with Astro and TypeScript.
An investigation of how data augmentation can be used to defend against adversarial attacks.
Using probabilistic LDA to extract and analyse topics in transcript data.
A collection of time series models used to analyse and model different problems.
Learning
A snapshot of what I have been exploring and leveling up in.
2026 February
Really enjoyed diving into Astro. Great framework for content driven static sites. Learned a lot with some help from Codex.
2026 February
Deep dive into agents and skills, and how to use hand-offs and subagents. This stuff will be huge.
2026 January
Replacing MongoDB with Azure Database for PostgreSQL. Created a DB and used alembic for migrations. Learned about pgvector for vector search. I still prefer Qdrant.
Featured Writing
A running log of what has caught my interest.
Feb 10, 2026
Why shipping AI is mostly about guardrails, observability, and feedback loops.
Jan 18, 2026
A concise way to tune expensive AI pipelines when brute-force sweeps are too costly.
Contact
If you're building something interesting, I'd love to hear about it.