Beyond the Numbers: UT Student Builds a Future in Computational Research - Profile Fiona Romanoschi

Simulating the movement of a glacier can mean tracking tens of thousands of variables over time. For fifth-year University of Texas at Austin student Fiona Romanoschi, the challenge is figuring out how much of that complexity can be reduced while still capturing the system's essential behavior. In her research at the Oden Institute for Computational Engineering and Sciences, she develops smaller mathematical models that can approximate the dynamics of ice sheets at a fraction of the original computational cost.

As Fiona prepares to graduate from UT in Spring 2027 with degrees in Informatics and Mathematics, a minor in Scientific Computation and a certificate in Computer Science, she is completing undergraduate research projects and applying to Ph.D. programs in computational science and statistics. Her interests have developed through a variety of research experiences, from statistical forecasting of political violence to computational modeling of glaciers, along with the guidance of mentors across several institutions.

Growing up in the Dallas area as the daughter of immigrants, Fiona initially envisioned a future in medicine. During high school, she participated in research programs focused on medical applications and regenerative medicine, including New York University's GSTEM with the Courant Institute and Stanford’s STaRS programs. Those experiences introduced her to scientific research and encouraged her to think more seriously about the methods used to investigate scientific questions.

Her high school statistics teacher was particularly influential in shaping those interests. "She wouldn't let me come to office hours without a specific question," Fiona recalled. "She always reminded me that asking specific questions helps other people understand how to help you, but it also forces you to think more carefully about what exactly you don't understand."

Outside the classroom, Fiona was also drawn to educational videos exploring mathematics and computation, including the YouTube channel 3Blue1Brown and a documentary about AlphaGo, the artificial intelligence system developed by Google DeepMind that famously defeated a world champion in the game Go. Together with her early research experiences, these influences encouraged her to explore how mathematics and computation could be used to understand complicated problems.

That growing interest led her to UT Austin, where she initially pursued Informatics. As she became more involved in research, however, she recognized that she wanted a stronger mathematical foundation to understand the methods she was encountering. During her junior year, she added Mathematics as a second major, complementing her studies with coursework in computer science and scientific computation.

The transition was a steep climb. Although she had always felt confident in her mathematical abilities, she discovered that proof-based mathematics required a different kind of understanding than the formula-driven problems she was accustomed to. Rather than simply knowing how to apply a method, she had to learn how to establish why it worked. Adding the major later in college also meant completing much of her advanced mathematics coursework in a compressed period. Over time, however, the experience changed how she approached both mathematics and research, encouraging her to focus more on underlying ideas rather than simply arriving at a correct result.

Throughout college, Fiona sought out opportunities to apply her growing interests to different research problems. She participated in two Research Experiences for Undergraduates (REU). The first at Worcester Polytechnic Institute, and the second REU at North Carolina State University.

While these experiences differed in focus, they helped Fiona discover a common interest in the mathematical and statistical methods underlying scientific predictions. They also exposed her to different approaches to research and to mentors whose perspectives helped shape her academic direction.

Her interests eventually led her to the Oden Institute, where she began working with Institute Director Karen Willcox and then-postdoctoral fellow Nicole Aretz on computational glaciology.

"My initial project focused on developing a reduced-order model of the Greenland Ice Sheet. Ice-sheet simulations use mathematical equations to describe how ice thickness and movement change over time, but high-fidelity models can be computationally expensive to run, particularly when scientists want to explore different physical conditions or future scenarios," said Fiona.

Reduced-order modeling offers a way to make these simulations more manageable. Rather than tracking every variable in the original model, the approach identifies a smaller representation that captures its most important patterns. Fiona's work also uses a method called operator inference, which learns the mathematical equations governing the evolution of those reduced coordinates directly from simulation data.

In 2025, Fiona received the Oden Institute's Ivo and Renata Babuška Scholarship for undergraduates in recognition of her research with the Willcox Research Group.

She subsequently applied to and was selected for the Institute's Moncrief Summer Internship Program in 2025. The internship allowed her to pursue computational science research as a standalone project and spend the summer developing her work further. During the internship, Fiona expanded her research to Pine Island Glacier in West Antarctica, a more complex modeling problem involving both ice thickness and velocity.

"The speedup we can achieve through reduced-order modeling is immense," Fiona said. "Sometimes our models run hundreds of times faster than the original simulations, which opens up the possibility of exploring scenarios that would otherwise be too computationally expensive to investigate."

But reducing computational cost is only part of the challenge. A smaller model must still reproduce the important behavior of the original system, and determining which information can be discarded without compromising its usefulness requires careful mathematical and computational work.

Since completing the internship, Fiona has continued working with the Willcox Research Group, investigating more structured approaches to operator inference that account for interactions between different physical variables. The research has encouraged her to think about broader questions involving mathematical models, and she is increasingly interested in how uncertainty, incomplete observations and imperfect modeling assumptions affect the reliability of scientific predictions.

In 2026, Fiona's research earned national recognition. She received two nationally competitive undergraduate research awards: the Barry Goldwater Scholarship and the Astronaut Scholarship. She was one of two UT students selected for the Goldwater Scholarship and the university's sole Astronaut Scholarship recipient.

Both scholarships have provided significant financial support during her undergraduate education. Just as meaningful to Fiona, however, have been the communities of scientists and fellow scholars she has joined through the awards.

Through the Astronaut Scholarship program, Fiona attended a gathering of recipients in Houston, where she had the opportunity to meet astronauts, established researchers and students pursuing scientific work across a variety of disciplines.

"I'm honored to be part of both scholarship programs. Beyond the financial support, I've especially appreciated meeting other scholars who are just as excited about their work and having a community of people I can turn to for advice."

Particularly important throughout that process has been mentorship. Fiona credits Aretz, Willcox and the rest of the Willcox Research Group with helping her grow as a researcher. "The mentorship opportunities I’ve had at the Oden Institute, and through my various certificate and REU programs have been very formational," she added. But beyond the technical guidance, she has also found inspiration in seeing her mentors pursue different paths within their own careers.

"Watching Nicole become a professor and Karen take on leadership roles connecting research with national policy has broadened my understanding of what a career in science can look like, and how I might eventually contribute beyond my own research," Fiona reflected.

The value Fiona places on education also has roots in her upbringing. "My family emphasized that while circumstances in life can change, the knowledge gained through education is something you can carry with you," she stated. It's a perspective that has stayed with her throughout college, as she has come to appreciate learning not simply as a means of earning a degree, but as a continual process of building knowledge and understanding.

Fiona's interests have also extended beyond her academic work. At UT, she spent two years with the fencing club, participated in several student organizations focused on data science, and served on the Natural Sciences Student Council. Her extracurricular pursuits have also taken her abroad, including a fully funded cultural and language exchange program in South Korea through the National Security Language Initiative for Youth (NSLI-Y).

As she looks toward graduate school, Fiona hopes to continue studying the mathematical foundations of scientific modeling, particularly statistical inference for complex dynamical systems. She aims to apply these ideas to physical and scientific systems, including climate-related applications, and eventually pursue a research career in an interdisciplinary environment such as a national laboratory.

"For graduate school, I'm excited about being in a new environment and learning from new mentors. Every mentor I've had has introduced me to different ways of thinking and approaching research, and I'm looking forward to discovering methods and perspectives I haven't encountered yet," Fiona said.