The University of Wisconsin-Madison campus. If you like lakes, you should apply. Source: University Housing
I am an incoming assistant professor at Wisconsin-Madison’s Computer Sciences department (starting Fall 2026), and my research focuses on natural language processing, computational social science, and responsible AI. I wrote the following so you can learn more about me and where I’m at, research-wise and department-wise.
Before applying, you do not have to reach out to me in advance, but you may do so if you’d like. I may not be able to respond to every email that reaches me, but I do read them. Please apply by our December deadline, and put my name down in your application.
If you are a prospective postdoc, summer research intern, undergraduate research assistant, or high school **research assistant, I sadly don’t have any openings at this time.
✨ It’s still super early for the 2025-2026 application cycle! This page will be updated over time as my thoughts evolve.
What is Lucy looking for?
Successful PhD students come in many forms. Most are driven by passion, from which they find resilience — a much-needed trait for the challenges of grad school. To be best prepared for the research we’d do, you would likely have a degree in computer science or a related field, or similar industry experience. You should also have some past experience demonstrating a sincere interest in some social science or humanities discipline, and/or an appreciation for language and linguistics. Producing good research and forming good research *taste* are two distinct but equally important skills, and your application should hopefully suggest the potential for both.
What might we work on?
My research group will bring together two research areas: human-centered AI development and computational social science. I don’t expect every student to span both, as each is a large, complex research area, and there is plenty of depth to achieve within a single side. Both sides benefit from knowing how models work and how they could be used, and an interchange of concepts and ideas between them makes everyone stronger. Both of these research areas benefit from interdisciplinarity, but I will likely train you to speak towards these communities from an NLP perspective. It’s good to define your home community early on, even if you do wish to be interdisciplinary, as it offers context for how you approach problems and guides your focus during your PhD.
With human-centered AI, my PhD students may generally wish to examine how social characteristics, behaviors, and interactions relate to models’ data ecosystems and development pipelines. Though much of my past work focuses on fairness, I’m also interested in studying the societal impact and safety of language models (LMs) more broadly. My research toolkit mostly consists of computational methods, but I am open to blending qualitative methods into this work. Some big questions we may try to tackle include:
Speaking of computational social science, my PhD students may also wish to develop NLP approaches for answering research questions relevant to sociology, psychology, education, and/or media studies. My flavor of computational social science overlaps with cultural analytics and computational humanities. When I work with data drawn from books and other media, I lean towards questions that center social aspects of that data. I also tend prefer projects that nudge us to be methodologically creative and/or connect language to insights around human behavior (my undergraduate degree was in cognitive science, and hints of that still remain!). Currently, I am intent on interrogating how LMs may be useful for researchers in the aforementioned disciplines, while considering practical constraints that may arise (e.g. resource-efficiency, subjective interpretation).
I understand that when you’re applying to PhD programs, you may reuse your statement of purpose across different programs. So, it’s totally okay if not all of the above in perfectly reflected in your statement, and when in doubt, you should apply anyways. If you do end up working with me, your interests will help me evolve mine.
Generally, I hope our work together will be thoughtful, timely, and impactful. Our research should go beyond **providing surface-level answers to questions like "Can LMs be prompted to do [some task]?" or "Are LMs biased against [some social group]?" Good AI research should also avoid being overly model-specific, and instead concentrate on something conceptual, to lengthen the potential lifespan of a paper. Aside from conceptual takeaways, our work may also produce artifacts such as code, data, and tooling that could be readily used by others.
A computer render of Morgridge Hall, which starting in Fall 2025, will be the home of the School of Computer, Data & Information Sciences. If you like new buildings, you should apply. Source: LMN Architects