Welcome! I'm Preethi and I am a researcher interested in Responsible AI, safety, fairness, and societal impact. My research focuses on evaluation, characterizing model behavior with respect to training data, and studying real-world harms in AI systems. I care deeply about designing and developing systems that are reliable, safe, and inclusive, and creating more robust and trustworthy evaluations of AI capabilities and performance.
I am currently wrapping up my PhD at UC Irvine, advised by Sameer Singh, and have been fortunate to be largely supported by an HPI Fellowship. I am also interning part-time on the Safety Team at Cohere with Seraphina Goldfarb-Tarrant, where I work on evaluating allocational fairness in enterprise settings such as hiring and investigating the validity of agentic evaluations.
Before starting my PhD, I worked as a Data Scientist at UnifyID and did my undergrad at Harvey Mudd College. Outside of research, I enjoy traveling, taking long casual strolls, and watching the NBA. I'm always on a quest for good eats, and I've recently rediscovered my love for amateur crafting.