Hands-on
I write code, design architectures, and debug production systems. I don't delegate the hard parts.
Founder, Encount AS · Inventor of Fusion (Citrus Analytics)
I build AI products, data systems, and production software for teams tackling technically difficult work. I've been doing this since 2010, and the systems I've built are still running.
My research spans the foundations of what we now call modern AI — and goes back further than most expect. In 1999, I built an intelligent web agent that used vector models and document similarity to learn each user's interests and surface relevant content — the same retrieval principles that drive today's AI search and recommendation systems. My Master's thesis at NTNU applied bio-inspired computing to cybersecurity: an artificial immune system that evolved its own pattern detectors to catch network intrusions. My PhD went deeper into language — building systems that decompose Norwegian compound words, map meaning across languages through semantic networks, and answer open-domain questions by extracting knowledge from text.
These are the problems at the heart of modern AI: how machines retrieve, adapt, and understand human language. I hold a PhD in Computer Science from NTNU, and I've been working on these problems for over 25 years. Today I apply that foundation to agentic AI systems, statistical analysis pipelines, and complex data platforms.
At Citrus Analytics, I invented and built Fusion — the AI-powered survey analytics platform the company offers. Building it made me a central part of building the company itself. Before that, I built the automated phone survey system behind Norsk Koronamonitor in three days, data pipelines for transport authorities, and a research platform that's been in production for 16 years.
I write code, design architectures, and debug production systems. I don't delegate the hard parts.
I take on problems that are technically interesting and genuinely matter — and say no to work that doesn't benefit from what I bring.
I bring in senior specialists when the problem truly benefits from it, not to fill seats.
I bias toward clarity, maintainability, and delivery. The right architecture today saves years of pain later.
Send a short note with the context, constraints, and what makes it hard.
Get in touch