Experiences

solving olympiad math and coding questions to train and fine tune models at scale ai.

researched and built large-scale multi agent systems from scratch at isara labs.
more to come soon :).

working with dr. gunawardena to build smartslides and codebench.
used by 1,500+ students. also tinkering with vlms.

led the ai policy team as a research scientist at humun.

cofounded param to build voice agents for clinical insurance workflows.
received $25,000 as part of the telora fellowship (~1% acceptance).

nj | sf

footnotes