hi, i'm shreyan.

• i study how multi-agent llm systems fail, mostly by attacking them. recently wrote a paper: when the judge intervenes.
• i enjoy building products. i made a consumer social app called ice, and have worked in product at two early-stage startups.
• i'm curious about what interfaces for thinking with ai should look like. i recently built essay atlas, an llm-drawn map of where my essays connect and contradict.
• i also write about design, media theory, and ai on my substack.