Shreya Havaldar

Shreya Havaldar

shreyah at seas.upenn.edu

About me

Hello!! I’m Shreya, a 5th year PhD student at the University of Pennsylvania working with Penn NLP. I am advised by Lyle Ungar and Eric Wong. My research is supported by the NSF GRFP. Previously, I attended USC (Fight On!!) where I received a B.S. in Computer Science & Applied Mathematics and worked with Morteza Dehghani at the Morality and Language Lab. I’ve also interned at Google Deepmind and Spotify Research.

I research sociotechnical alignment of LLMs, with a focus on evaluating and mitigating cultural bias. My work looks at how LLMs handle context-dependent phenomena such as implied meaning and social norms, and I have developed frameworks to assess how they represent subjective concepts like politeness, emotion, and individualism. I have also built methods to improve model outputs, including preserving speaker intent in translation and adapting advice to multicultural users.

I’m currently on the postdoc market! I’m interested in pluralistic alignment in LLMs.

Beyond my research, I care deeply about representation and accessibility in ML. I founded the UPenn chapter of Women in ML (WiML) in 2023, and I’m currently creating and teaching a new course on designing LLM agents for Penn undergrads who are beginners in CS.

I also like to cook, run, and ask strangers if I can pet their dogs. Please reach out if you want to chat about my research, the GRFP, or the PhD application process!

Recent News!!