I was the second hire at Letta, a research company focused on memory for AI agents. I contributed to and lead several efforts:

• • Building a stateful agent API to support a million agents [blog]
• • Giving agents the ability to search over large file systems in the cloud [blog]
• • Designed Letta Evals OSS, an evals package designed for stateful agents [blog]
• • Designing benchmarks like Context-Bench [blog] [post]

Before that, I was an engineer at Databricks, working on systems and deployment. During this time, I also was working on generative 3D models at Stanford.

I studied at Berkeley.

X•Instagram•Email