Brihi Joshi

I am a final year CS PhD Candidate at the University of Southern California, advised by Xiang Ren and Swabha Swayamdipta as a part of the NLP Group, INK Lab and Dill Lab. I also work closely with Tim Paek and Rik Koncel-Kedziorski at Apple on various topics in personalization.

My research focuses on making AI systems genuinely usable for diverse people. I study data and methods that explicitly optimize for human–AI ecosystems, and I develop human-grounded evaluations that are scalable and actionable. During my PhD, I have pursued two complementary directions toward this goal:

• Personalization: I evaluate and improve how AI systems adapt to different users. My work explores new interaction paradigms [Ongoing Works], including multi-turn aware [EMNLP’25c] and personalized AI systems [ACL’25], to better model users’ backgrounds, needs, and preferences [EMNLP’25a, EMNLP’25b].
• Interacting with AI systems: I studied how AI systems and people can communicate better for the betterment of both. I investigate how models can produce high-quality explanations [EMNLP’22, ICLR’24] and how such systems can improve understanding, trust, and decision-making for lay users [ACL’23,Under Review] and domain experts [EMNLP’24, Outstanding Paper Award].

My research is supported by the USC Annenberg, 2023 Amazon ML and 2024 Apple Scholars in AI/ML Fellowships.