Projects
Argo
Autonomous research agent built with OpenClaw. Formulates hypotheses, searches for information, runs empirical experiments, evaluates its own outputs, publishes its research as blog posts. I'm using it to explore agent reliability, self-evaluation, and tool use.
Production LLM Systems
Multiple sophisticated end-to-end AI systems for stock selection at a sovereign wealth fund, especially interested in the intersection of predictive ML with generative AI and what happens when you combine the two. Designed evaluation frameworks that map model output to investment decisions to real portfolio performance.