Marko Cvjetko

PhD student · Inria, Flowers Team · University of Bordeaux

email github scholar cv demos

Hi 👋, I am a PhD student in the Flowers AI & CogSci Lab at Inria Bordeaux, 'supervised by Pierre-Yves Oudeyer and Clément Moulin-Frier. I develop curiosity-driven AI methods for assisted scientific discovery, and I deploy these within Artificial Life and biologically grounded substrates to study open-ended evolution and self-organization.

Previously, I worked at Haesler Lab, where I developed computer vision tools to support research of the neural circuits behind intrinsically motivated behavior in mice.

AI for Science · Reinforcement Learning · LLMs · Artificial Life

Research

Complex systems give rise to striking phenomena that we still cannot reliably discover, characterize, or control. A central aim of my thesis is to build AI tools that autonomously navigate vast spaces of possible simulations to surface phenomena that matter to human researchers. To this end, I draw on and adapt methods from reinforcement learning, generative and foundation models, and evolutionary algorithms.

Beyond building these tools, I am drawn to the questions they uncover: what are the minimal conditions for open-ended evolution, the process by which simple interactions generate increasingly complex forms? What do concepts like memory, agency, and intelligence mean in systems where complexity arises from the bottom up?

I strive to bring my methods to tasks other researchers find meaningful. To this end, I have a fruitful, ongoing collaboration with the Levin Lab at Tufts.

Publications

Conference papers

Workshop papers

Late-breaking abstracts

Service & supervision

News