On a fall afternoon, at NRG Systems’ headquarters in Hinesburg, Vt., University of Vermont engineering student Ethan Perry ‘26 is giving a half-hour talk to a group of the company’s engineers and scientists. On a screen glows an image of a device that looks like a Mars lander in the middle of a farm field. It’s one of the company’s Flare SRA systems—basically, a super-accurate tool for measuring how much sunlight is falling onto a location, and a bunch of other variables, including windspeed, temperature, and rainfall.

Since the early 1980s, NRG has been a global leader in creating and deploying measurement tools for wind and solar energy companies. The Flare tower is trusted by major electricity utilities and developers for making multi-million-dollar decisions about where to build solar arrays and how profitable those projects could be. 

As a research intern, Perry spent the summer exploring the system’s maintenance photos and other data, looking for ways to improve the quality of the information each tower, and its maintenance team, collects. Now he’s explaining how his research could help NRG push their advanced system closer to perfection.

“As a reminder,” he begins, pointing to two gleaming devices that look like tiny UFOs bolted to the top of the tower, “Each Flare SRA system come with a pair of pyranometers.” These are the sensors that measure the amount of sunlight energy that is coming down—what the engineers call solar irradiance. Having two of these sensors is redundant—and that’s the point. If they both have the same reading, it gives users confidence that the data are accurate. 

Perry flips to another slide showing two side-by-side solar panels attached to the tower—one labeled “clean,” the other “dirty.” These allow a customer to compare power outputs, “as the clean panel is kept clean and the dirty panel accumulates dirt, pollen and whatever else,” Perry says. This data reveals how often a company would have to clean their panels at that site. And, each week, technicians measure the levelness of the panels—to make sure the two panels are facing the sun in exactly the same way.

The tower technology is sophisticated, elegant, and industry-leading—but the sensors, and the technicians that maintain it, collect a lot of data that comes in different forms. “It's kind of scrambled,” Perry says. To help improve the data and its ease of use, Ethan worked under the supervision of NRG data scientist Alexandra Arntsen to build a digital processing model that can unify each site visit’s photos, technician forms, and metadata into a single machine-readable structure. This opens doors to automated checks: assessing whether technicians took all 19 required images, verifying sensor orientation, and diagnosing anomalies in irradiance readings.

“Using the digital data processing model that I developed, we can combine all this data into one cohesive package,” Perry tells the team of scientists. In other words, his work can make the field data easier for computers to interpret, easier to analyze at scale, and easier for NRG’s clients to review. 

Toward the sun

Moments like this—an undergraduate given an opportunity to conduct independent research, improving a world-class measurement system—don’t just happen. Ethan’s path reflects the blend of curiosity, rigor, and opportunity that UVM cultivates.

He started his UVM career as a computer science major, the kind of student who began coding when he was a kid and taught himself the coding language Python in early high school. “It’s something I love,” he says. “For many years, I’ve been doing automation tasks for fun with my buddies.”  But as he talked with his professors and thought about his future, something clicked. “Waking up and doing something that I’m really proud of is important to me,” he says. “Solar, wind—renewables in general—felt like the direction I wanted to go.” 

That calling led him to switch to electrical engineering, and eventually to apply for and receive support through the Richard Barrett Scholars Program, which places UVM students in sustainability-focused internships and funds their research. The Barrett program usually connects UVM students with professors, working in on-campus labs. But the program connected Perry with NRG, the first time that a Barrett Scholar has worked with an industry partner. And NRG, with its mix of advanced data science and mission-driven engineering, turned out to be a great fit.

“It’s unique compared to just working with professors,” Perry says. His work with data-model and image-processing improvements for NRG is not just a training exercise. What he and Arntsen have developed can be put to use in the company’s current operations. “We’re doing something that’s going to be useful now,” he says.

Energy Investments

Professor Luis Espinosa, one of Perry’s key mentors, describes him as a standout student in some of UVM’s most demanding courses—including a foundational class on signals and systems as well as advanced “hardcore theoretical courses,” he says. “He was happy to learn that abstraction pays off,” Espinosa says. “He saw beyond formulas. Once students realize abstraction can take you farther, that deeper understanding lets you provide guarantees”—like guaranteeing the accuracy and ease-of-use of a complex data stream coming from an NRG tower. And guarantees are what utility clients expect when they invest millions in renewable infrastructure. 

Espinosa also remembers helping Perry navigate a project that required experimental design, careful reasoning, and coding discipline. “You can search as much as you want on the internet,” Espinosa told him, “but if you don’t put pen and paper together and test something, you won’t get to a solution.” 

That practical problem-solving rigor has become one of Ethan’s strengths in his work at NRG, says his research supervisor Alexandra Arntsen, a UVM alumnus from the class of 2012 and data scientist with a PhD from Dartmouth. She supported Perry in exploring questions, “where we don’t know the pathway or the process or the answer,” she says. In contrast to a typical engineering intern, who might test existing systems or products, Perry developed machine-learning and computer-vision approaches through experimentation and self-directed research. “It’s been like working with a first-year PhD student,” she says. Arntsen says Ethan’s analysis strengthened the company’s confidence in a new app it’s developing for its technicians to use in the field. “Ethan’s work validates our concept,” she says. “And he’s informing us of what picture concepts are important if we’re going to execute this model.”

For Perry, this summer confirmed both his skills and his ambitions. His paid internship is continuing this fall semester, part-time, while he takes his final courses at UVM. And he hopes to stay in renewable energy—maybe even at NRG. “I’d love to get a full-time job here,” he says. And for NRG, investing in Perry and other interns represents something meaningful too: not just improving their products, but also supporting the next generation of engineers who could make the path to clean energy easier and more profitable.