Alish Dipani
I study the visual perception and metacognition of AI models and agents, using human behavior and the brain as a reference.
Actively looking for research internships in Summer and/or Fall 2027.
Neuro & cognitive science for AI
Probing AI models and agents with controlled experiments from psychophysics and metacognition, and comparing their internal representations with the human brain.
AI for neuro & cognitive science
Building AI models that predict human brain activity and behavior, as well as agents that conduct in-silico experiments.
Lately, I've also been expanding into technical AI safety through Georgia Tech's AI Safety Initiative, out of concern about the risks. (why?)
I'm a NeuroAI PhD student at Georgia Tech, advised by Apurva Ratan Murty and Dobromir Rahnev, supported by the CoCo Fellowship from the Center of Excellence in Computational Cognition. My background is in computer science, vision science, and computational cognitive neuroscience. Previously, I did research at MIT Brain and Cognitive Sciences, Northeastern University, and CNRS, and worked as a machine learning engineer at Baylor College of Medicine. I also teach and create content for Neuromatch Academy, and contribute to open-source software.