Hello! I'm Hari, a senior undergradute.

I'm primarily interested in using Bayesian models of cognition to understand constrained decision-making in humans and LMs, particularly within cooperative environments. More recently, I've been motivated by problems in technical AI safety and alignment. I sometimes write about these thoughts.

Currently, I'm an IRIS Fellow with the CoSi Lab at NUS, advised by Prof. Tan Zhi Xuan. My work aims to improve agentic assistance to boundedly-rational users.

I also study the effect of near-misses in controllable environments, advised by Prof. Nandan Sudarsanam at WSAI, IIT Madras. Previously, I worked on an IIT Madras-UIUC project, analyzing models for the early detection of Dementia in Alzheimer's patients.

In my free time, I debate competitively, play chess, and take photos.

Elsewhere

• Email hariganeshtangirala [at] gmail [dot] com
• LinkedIn /in/hariganeshtangirala
• Twitter @bayesianhouse