I am an applied scientist at AWS AI, where I work on making LLM agents efficient: adaptive execution for Amazon Quick, post-training compact models to speed it up, and cost-aware agentic planning. I am also a visiting researcher at MIT CSAIL, where I co-lead, with Mike Stonebraker, research on AI agents for enterprise data.
My research builds data-centric AI systems that automate and augment complex user and system work. This research sits at the intersection of AI, data systems, and human-computer interaction, and has produced an extensive body of publications, including award-winning papers at premier venues, as well as systems deployed in production. Since 2023, I have focused on agentic AI systems for enterprises, their efficiency, reliability, and evaluation through realistic benchmarks.
Previously, I was Chief Research Scientist at Sigma Computing, where I set the AI roadmap and built the research group from scratch. I also held research positions at Megagon Labs and IBM Research. I was a postdoc at Stanford and a member of IDL at the University of Washington, and I received my Ph.D. in computer science from Brown University, with a thesis on computational brain connectivity using diffusion MRI. I also co-founded Fitnescity to provide easy access and data analytics for lab tests.