
Founded India’s first Alignment Research Fellowship: 600+ applicants across 40 STEM universities, 24 selected (top 4%), 10 papers published or under review.
Project Tags
India produces some of the world’s sharpest CS, Physics, and Mathematics graduates. But in 2022, there was no structured pathway for any of them to enter AI safety research. No fellowship. No reading groups. No community. The talent existed—scattered across 40+ STEM universities—with no connective tissue.
Labs abroad were hiring from the same handful of Western graduate programmes. An entire subcontinent of potential researchers was invisible to the field.
Founded India’s first Alignment Research Fellowship. Built a pipeline from scratch: 600+ applicants across 40 STEM universities, evaluated on algorithms, ML theory, and research aptitude. Selected 24 fellows—a 4% acceptance rate—and placed them in structured research tracks covering interpretability, LLM robustness, scalable oversight, and model evaluations.
10 papers now published or under review. 5 fellows pursuing research independently or at labs like the Mila-Quebec AI Institute. The initiative was acquired by SteadRise and scaled globally.