top of page
bay area final.png


My research develops computational models of communication, with an emphasis on the real-world consequences of speech acts.

For example, how will an utterance affect a listener's beliefs and actions?  And how do people (or LLMs) weigh these outcomes? Answering these questions can both further our understanding of humans and ensure the deployment of LLMs goes well for society. 

I'm currently working on LLM safety at Anthropic. Previously, I got a PhD in cognitive science (advised by Tom Griffiths at Princeton, and supported by a NDSEG fellowship; with an internship at DeepMind). Before the PhD, I was a data scientist and engineering manager at Automatic Labs (2013-2014) and Uber (2014-2019).

sumers_headshot (1).jpg
bottom of page