Nupoor Gandhi (she/her)

I am a PhD student in Machine Learning and Public Policy at Carnegie Mellon University advised by Emma Strubell. Previously, I received my bachelors in Computer Science with a minor in Math at the University of Illinois at Urbana-Champaign. At Illinois I was a part of the Text Information Management and Analysis Group and the Health and Social Media Group working with ChengXiang Zhai and Dolores Albarracin. I was also a Data Science for Social Good Fellow at Imperial College London.

I develop methods to make large-scale text analysis accessible. In particular, I am focused on reducing the high cost of structured information extraction, particularly in settings where large-scale expert annotation is prohibitively expensive. A motivating example of my work is in the inacessibility of climate policy information. Much of what governments commit to, let alone implement, is buried in long, dense, heterogeneous documents that are difficult to systemically analyze: Robust structured information extraction can make climate policy legible and measurable at scale.

My thesis is focused on structure as a core modality for efficient, scalable knowledge work. Structured representations are richly expressive, compact, and reliably grounded in text. Crucially, structure is decomposable by definition, making it possible to distribute annotation efforts across heterogeneous groups, including both language models and human annotators.

Contact: nmgandhi(at)cs.cmu.edu Google Scholar CV