about

after a struggle of 5 years, i graduated from bits pilani with a bachelors in electronics and a masters in mathematics. i've spent the last ~3 years as a software engineer in the startup ecosystem.

started as an early engineer at houseware back in late 2023, building agentic systems when agents were just an early concept in research papers. had a lucky exit as the startup got acquired, after which i spent almost a year and a half at launchdarkly learning how to build backend systems for extreme scale. then a short month with conscious engines, an ai lab working on proactive personal intelligence, after which i took the plunge and left the safety of my job to go full time into ai/robotics research.

my current focus is uncertainty quantification for robot policies, or in plain words, predicting and preventing failures in a deployed policy at inference time.

the bigger problem i care about is self-improving robotic systems, robots that get better the more they run. the spaces i'm exploring sit along that closed loop: failure and uncertainty detection, fault attribution, automated reward modelling, and leveraging world models for policy improvement.

some things i deeply believe in

• synthetic humanoids
• continual, omnipresent personal intelligence across consumer software and hardware
• bio-enhanced humans

when i'm not building software or getting existential about the future, you can find me planning my next backpacking trip, breaking my legs training for a full marathon by the end of 2026, or getting back to singing.

feel free to reach out on twitter or linkedin.

oh, and here's my github and resume (no promises of resume being updated though).