Jake Snell

Also known as Jake C. Snell

Jake Snell is an associate research scholar in the Department of Computer Science at Princeton University, where he works with Tom Griffiths and develops deep learning algorithms informed by probabilistic models. After earning a biomedical engineering degree at Yale University in 2010, he completed a master's degree in 2014 and a computer science PhD in 2021 under Richard Zemel at the University of Toronto. With Kevin Swersky and Zemel, he introduced Prototypical Networks for few-shot learning at NIPS 2017, learning a metric space that classifies examples by computing distances to class prototypes. He received a DataX Postdoctoral Fellowship at Princeton in 2023, presented a metalearned sequence model for nonparametric Bayesian inference at NeurIPS 2024, and won an ICML 2025 Outstanding Paper Award with Griffiths for "Conformal Prediction as Bayesian Quadrature." In March 2026, he announced that he will join the University of Cambridge's Department of Engineering as an assistant professor starting in January 2027.

Now at: Princeton University
Badges: Bridge
Boomerang

Timeline

2010–2014

Yale and a Toronto master's

He earns a B.Sc. cum laude in biomedical engineering at Yale University in 2010. In 2014, he completes an M.Sc. in computer science at the University of Toronto.
2016–2018

Research internships in industry

He works as a research intern at Twitter, now X Corp., in 2016. In 2018, he takes a research internship at SK Telecom.
2017

Prototypical Networks

With Kevin Swersky and Richard Zemel, he presents "Prototypical Networks for Few-shot Learning" at NIPS 2017. Prototypical Networks learn a metric space in which classification is performed by computing distances to prototype representations of each class.
2020–2021

A Toronto PhD with Richard Zemel

In 2020, he and Richard Zemel propose a Gaussian process classifier for Bayesian few-shot classification based on a combination of Pólya-Gamma augmentation and the one-vs-each softmax approximation. He completes a PhD in computer science at the University of Toronto in 2021 with the thesis "Learning to Build Probabilistic Models with Limited Data," advised by Zemel.
2023

DataX fellow at Princeton

He receives a DataX Postdoctoral Fellowship at Princeton University and a TMLR Expert Reviewer Certification in 2023. As a DataX postdoctoral researcher in Princeton's Department of Computer Science, he works with Tom Griffiths and develops deep learning algorithms by drawing insights from probabilistic models.
2024

Metalearned neural circuits at NeurIPS

At NeurIPS 2024, he, Gianluca Bencomo and Thomas Griffiths present "A Metalearned Neural Circuit for Nonparametric Bayesian Inference." By simulating data with a nonparametric Bayesian prior, the method metalearns a sequence model that performs inference over an unlimited set of classes.
2025

An ICML Outstanding Paper Award

In April 2025, his coauthored paper "Binary Climate Data Heightens Perceived Impact of Climate Change" appears in Nature Human Behaviour with Grace Liu, Thomas Griffiths and Rachit Dubey. With Griffiths, he also writes "Conformal Prediction as Bayesian Quadrature" in 2025, revisiting conformal prediction from a Bayesian perspective. The paper wins an ICML 2025 Outstanding Paper Award, one of six chosen from 3,260 accepted and 12,107 submitted papers.
As of October 2026

Associate research scholar at Princeton

He is an associate research scholar in Princeton's Department of Computer Science, where he works with Tom Griffiths. In January 2026, he organizes a workshop on Principled Design for Trustworthy AI at ICLR 2026. In March 2026, he announces that he will join the University of Cambridge as an assistant professor in the Department of Engineering starting in January 2027.