I spend a lot of time thinking about brains, learning, and the implications of technology. Sometimes those thoughts become research, projects, and writing.
I’m a PhD student in NeuroAI at Yale, working at the intersection of causal inference, manifold learning, and theoretical neuroscience. My research asks: how can we use generative models of neural data to uncover the mechanisms that guide the brain?
I studied computer science and pure mathematics at Columbia University, where I did research on manifold inference.
I use mathematics and computational tools to study brains and artificial systems. I’m interested in how they learn and change over time, and what their structure can tell us about how they work.
Selected works
Two neural systems can look alike on standard similarity measures while organizing information very differently. We show why matching their outputs is only part of the story, and how studying the organization of neurons reveals what those measures miss.
Read the paper
Can recordings of brain activity tell us which neurons influence one another? We use diffusion models to infer candidate connections and their timing, then compare them with known wiring in the nervous system of C. elegans.
Read the paper
A model can predict brain activity without working like a brain. We look inside a neural foundation model to see how it processes visual information, finding both parallels with biology and differences hidden by prediction accuracy alone.
Read the paper
A tool for checking whether changes to an AI system still meet the standards people expect of it. Start with examples, make your judgments, and see how a candidate model measures up.
Explore the demo
An interactive tool for exploring how groups of neurons respond over time, using visualizations of recorded brain activity.
View on GitHub
I write on Substack about technology, society, and the ideas I keep coming back to.