I work on latent variable discovery in high-dimensional neural data, using neural networks and variational autoencoders. Specifically, my work involves incorporating informative priors in what's otherwise considered "black box"-models, so that we can infer understandable and meaningful connections in our brain.