Joey David

Intelligence has fascinated me long before I was introduced to AI/ML. I looked for it in friends, in adults, even in animals. As a teenager, getting into compsci and math, it became clear that my previous conception of it was limited.

Rather than being only an aspect of biological brains; intelligence is a measure of a system's ability (nearly any system, not just conscious ones) to find and exploit structure in problem spaces, in abstract concepts, and in the physical world.
I think that for instance, life evolving from simple amino acid chains to the complexity of the human brain seems much more elegant and coherent when intelligence means structure-finding. Even on a constant fight against entropy (I strongly recommend A Brief History of Intelligence for a good holistic view), this system-oriented view is much more intuitive. So is the wave of algorithmic progress made since the 1950s.

Since then, I've kept working on it! As of early 2026, I am completing my M.Sc. in AI/ML at PSL University (under ENS, Mines Paris, and Dauphine) and working as a research intern with the MILeS team at LAMSADE. My current research is on the information-theoretic limits of reasoning processes in quantized LLMs, with the broader question of what problem-solving information survives aggressive adaptation and compression.

Outside of formal research, I like to build my own tools, software and experiments. I'm always happy to chat about anything, do feel free to reach out.