RSEs a gamechanger for researchers

Building an RSE group at UD has been a goal of Rudi Eigenmann’s since he joined the faculty in 2017.

“Science needs RSEs because chemists, for example, need to concentrate on chemistry, and it takes so much just to learn the chemistry,” said Eigenmann, Distinguished Professor of Electrical and Computer Engineering and co-investigator on the grant. “With more and more work being done computationally, chemists also have needed to become computational experts. But with RSEs, chemists and other domain scientists don’t need to spend so much of their time trying to learn the computational skills as well.”

Eigenmann credits UD’s success in winning the grant to Chandrasekaran’s leadership and the strong collaborations among data science groups in high-performance computing (HPC) and artificial intelligence on the academic side, working with the UD Research Office and UD Information Technologies (IT), which support researchers with a variety of technology resources and services. 

“AI is such a large and encompassing juggernaut right now — you have to approach it from many different directions across research, teaching, learning and administration,” said John Huffman, director of IT-Research Infrastructure, who is also a co-investigator on the project.

As pointed out at a recent NSF meeting Huffman attended, a university’s research capabilities once were based on the size of its computational resources, but those days are likely waning.

“What really distinguishes institutions is having the expertise to develop software with RSEs as part of your research support,” he said. “The breakthroughs are in trying new methods and being able to squeeze out the performance from HPC systems. That’s the next evolution in how you grow as research leaders.” 

UD breaking ground with RSE program

UD is the first university funded by NSF SCIPE to create the RSE pipeline. Curriculum development is well underway for the program, which will be launched in 2026 and will be sustained with support from the Graduate College. Chandrasekaran and project coordinator Samantha Smith already have students lining up to pursue the training. 

Involvement in research projects is a key component, as Yamini Pravallika Medapati, Nihaal Chowdary Surpani, Aakanksh Chittiprolu and Kushagradheer “Kush” Srivastava, all master’s students in data science, well know. Chittiprolu and Srivastava are examining vastly different research topics on their RSE path. 

In August, Chittiprolu was assigned to UD’s baseball team. He’s not a player — cricket and tennis are more his thing — but he’s learning more about baseball as he analyzes data collected by a TrackMan digital scouting tool, ranging from a pitcher’s movement profile to the location, velocity and spin rate of every pitch thrown. 

“I’m using a programming language I know, called ‘R,’ to start building a dashboard that players and coaches can use to better understand where improvements could be made,” he said. “If a pitcher needs to improve his slider, the dashboard will show that.”

Meanwhile, Srivastava is working with Ben Bagozzi, professor of political science and international relations and co-investigator on the grant. 

Srivastava is extracting data in English and Russian, such as text and images from Russian government websites from 1998 to the present, to see if a pattern emerges in how the Russian government prioritizes and signals its foreign policy agendas. 

In addition to his adviser’s support, he’s working alongside Daria Blinova, a Ph.D. student and research assistant in the Department of Political Science and International Relations, while another student works nearby on a coastal science project. 

“We do get stuck,” he said, “but usually not for long. There is a lot of professor support and resource materials.”

Democratizing AI 

Given Delaware’s vulnerabilities to sea level rise and climate change as the flattest state in the nation, in a geological hotspot that is sinking, and under present-day conditions where the intensity and frequency of major storms is higher now than even 10 years ago, being able to predict coastal flooding and erosion is more critical than ever. 

Tian-Jian (Tom) Hsu, professor and director of UD’s Center for Applied Coastal Research, is part of the grant’s effort to show how the RSE framework can expand AI applications to new areas of coastal science research.

Currently, when Hsu’s team tries to model flooding in Delaware’s Inland Bays, it takes a week of computer run-time to see the impact, and the model cannot provide any finer resolution than 30 kilometers (18 miles). 

Two RSEs are now assigned to the center, working with Hsu and with Prof. Carolyn Voter, and are focused on improving predictions of flooding and water quality issues at the storm scale — the passage of a hurricane over a week-long period. 

“The RSE is so important for us,” Hsu said. “Using AI, you can make everything faster, you can reduce time — from a graduate student taking a week compared to an RSE taking one hour. With them, we can really maximize productivity. But first, we need to create a constant pipeline of RSEs and make this an official career because it is definitely needed.”