I’m currently a member of the technical staff at Anthropic.
Previously, I worked at Zetta AI, a startup that offers neural circuit reconstruction technology to research groups. I also ran a (somewhat) regular reading group in the EleutherAI discord server that discussed AI interpretability research with a focus on work relevant to AI existential safety. I often go by “Nick” (he/him/his).
Before that, I was a graduate student in the Seung Lab at Princeton University working on connectomics of tissue samples imaged by electron microscopy (EM). Most of my graduate work focused on synapse detection and assignment in EM volumes using convolutional networks. I built a software system for performing this task at scale, and our lab has also used the same system to perform automated segmentation of mitochondria and cell nuclei. I’ve also done some work on analyzing the connectivity patterns of cortical networks, and determining the cell types of ganglion cells in the mouse retina.
Before my graduate program, I did a research fellowship with Karen Berman at the NIMH using neuroimaging techniques. I analyzed diffusion-weighted imaging data to help characterize the white matter of children with Williams Syndrome, and analyzed gray matter volume distributions across typically developing volunteers.