About Me

I'm an AI safety researcher working on corrigibility with the goal of mitigating existential risk. Currently, I'm building a corrigibility benchmark and finetuning open source models to be more corrigible as part of the CAST framework.

I'm a philosopher by training. At the University of North Carolina, I worked on meta-ethics and evolutionary debunking arguments against moral realism. I have also worked in the philosophy of science and studied physics as an undergraduate.

I spent a year teaching at a middle school in the Marshall Islands, on a tiny island settled by displaced Bikinians.

I also design, pattern, and sew apparel.