Hi! I’m an A.I. researcher, academic, and entrepreneur, based in San Francisco.

• In 2026 I co-founded Elorian (with Andrew Dai and Yinfei Yang), where we are building frontier visual thinking models. We are backed with $55m from Menlo, Altimeter, Striker, Nvidia, and Jeff Dean.
• From 2021-2026 I was an assistant professor at Harvard (HBS, CS, Kempner) and PI of the SAFR AI Lab, where our research focused on problems in data-centric AI and AI safety. The research I’m most known for includes: the first SFT method for knowledge removal from models, the first algorithm with “provable” guarantees for learning an unbiased model with respect to exponentially many protected groups, the best known bounds connecting differential privacy and adaptive data analysis, and investigating the privacy properties of LLMs.
• From 2025-2026 I was on sabbatical at Google, working as a research scientist on the Gemini Data core team, and a co-author on Gemini 2.5.
• From 2016-2021 I was co-founder & CTO of Welligence, which is a leading energy data and forecasting company. After bootstrapping to 60+ FTEs, in 2024 we sold a minority stake in the business. Since then the business has continued to grow significantly, which I am proud to support as a board member.
• In 2020 I received my PhD in machine learning from UPenn, under the guidance of the legendary Michael Kearns and Aaron Roth. Prior to that, I got my undergrad degree at Harvard where I got a crash course in pure math through math 55, in startups as an intern at the device company WHOOP, and in squash as a benchwarmer on a perennial collegiate powerhouse. I also got to hang out with my twin brother which was pretty cool.

If you want to get in touch, I’m seth at elorian dot ai.