About me

I am a PhD student at NYU, advised by Prof. Danny Huang. I am also fortunate to get guidance from Prof. Jessica Staddon, and Prof. Rahul Chatterjee in various capacity.

Research interests

My research focuses on responsible and human-centered AI, where I reduce end-user security, privacy, and safety risks in LLMs, with particular attention to vulnerable users such as individuals experiencing technology-facilitated intimate partner abuse (TFIPA). I use empirical methods to uncover risks in generative AI applications, and design interventions that address these failures. My work contributes to safer AI systems and offers evidence-based insights for policy and practice.

Future research directions
 - Algorithmic auditing of “social AI”: Examining safety, security, and moderation challenges as generative AI is used to create content in social networks. Interested inmeasuring whether platform protections are effective and equitably applied across users of different social backgrounds and age groups.
 - Emperical measurements of human-centered or network security, privacy, and safety.
 - Emperical evaluation of LLMs security and privacy advice for diverse group of end-users, such as consumers, developers, and at-risk population.

Recent Updates

• January 2026, Our paper “Assessing LLM Response Quality in the Context of Technology-Facilitated Abuse” is accepted at USENIX 2026.
• December 2025, presented my work Learned, Lagged, LLM-splained: LLM Responses to End User Security Questions at ACSAC 2025.
• December 2025, presented my work Experimental Aspects of Evaluation of LLM Responses to End-user Security Questions in the LASER workshop cohosted at ACSAC.
• November 2025, gave talk at Cornell-Tech security seminar about my work on evaluation of LLMs ability to help TFA surviors with their technology questions and end-users with security questions.