AI is the strange loop of the modeler building a model of modeling itself.
AI researcher in Munich, building adaptive, embodied systems that keep learning after they ship. Underneath the work is a simple conviction: to understand anything is to model it — and intelligence, from a single cell to a whole society, is the act of modeling the world.
On-chip learning, Loihi 2
Sub-millisecond on-device learning at 6,600× lower energy than an edge GPU.
CLP — learning prototypes
Rehearsal-free continual learning, ~85% less catastrophic forgetting.
Predictive world models
Self-supervised, multi-time-scale continual learning on LIBERO.
Why frugal, embodied AI
Against brute-force AI; for systems that run on a chip in a field.
Multi-time-scale learning
Why embodied autonomy stalls between a working demo and a deployed system.
Through the lens
Film photography, travel, and coffee as small ongoing experiments.
The next frontier in embodied AI is not a larger pre-trained model, but learning that happens on many time scales at once — milliseconds to years.
I'm drawn to the brain from every angle: neuroscience, psychology, NeuroAI, and the philosophy of intelligence. The same instinct shows up away from the desk — photography, travel, and treating coffee as a small ongoing experiment. This site is meant to show the whole picture, not just the CV.