The hike was unlike anything Brett Storoe had ever done.
The Naperville native was used to hiking on flat terrain in Illinois with his family, but this time he was in Chile trudging up an ice stupa — a man-made glacier used to store water in the winter. While he was familiar with ice stupas by tracking them through a project he had been working on, he had never seen one in person.
“I’ve seen them in photos for over a year. I’ve been staring at them, but seeing them in person, how much they’ve grown and the actual size of them was incredible to see,” Storoe said. “I thought they were just maybe a couple people tall, not much taller than that. But seeing them in person, they were multiple stories.”
For that hike, he would need to trade in his athletic shoes for snow boots with spiked bottoms. Climbing the stupa was a dangerous endeavor, one where he had to use the spikes of his shoes to dig his own hole for grip. It was all worth it when he made it to the top, he said.
“It was just a great experience to connect back with the work that I’ve been participating in for the past year,” Storoe said.
Since fall 2024, Storoe has been developing an algorithm that determines where ice stupas can be built to address water scarcity in Chile. His work is part of a larger collaboration between the Milwaukee School of Engineering and Chilean-based startup Nilus, an organization dedicated to utilizing artificial intelligence to bring back ice glaciers.
Storoe became involved with the project through an AI club, which the 20-year-old college junior has been part of since his first year at the university.
“Growing up, I was very math focused. Those are the classes that I enjoyed. I liked the problem solving aspect of it,” he said.
An alum of Naperville’s Neuqua Valley, he started taking coding classes in high school, which helped solidify his interest in engineering. He had heard about the Milwaukee School of Engineering’s AI club while he was at Neuqua and knew he wanted to be part of a club like that.
For his first year at the AI club, Storoe worked on a project aimed at predicting food price spikes in developing countries to help people in those countries stock up before the cost increases hit, an experience that sparked a larger interest in the intersection of engineering and the environment.
Fast forward to fall 2024, a professor at the university’s business school connected students from the AI club to Nilus to work on an algorithm responsible for determining locations where artificial glaciers could be built.
“I knew absolutely nothing about anything related to Chile or artificial glaciers,” Storoe said, but that lack of familiarity also made the project exciting.
According to a news release, ice stupas are “formed by channeling water from higher altitudes through underground pipes, and then sprinklers spray water into subzero air to create these ice formations.” The stupas melt in the summer, providing water to communities during a time when water is scarce.
The algorithm created by the students uses weather data to “know when to open the sprinklers for optimal glacial growth,” the release said. Students also made “a customized Swarm Rag approach that allows meteorological experts to analyze changing weather data over years of large datasets,” the release said.
Storoe was primarily responsible for growth analysis and volume tracking. Initially, Nilus was using a manual process for tracking growth of the stupas, using images from static cameras set up in the mountains and comparing those images to determine how much a stupa had grown or shrunk.
“Previously, (Nilus) would just look at the images and kind of hand label them and say, ‘Oh, this day had a ton of growth, this day it shrunk a little, or this day it stayed still,’” Storoe said. “So it was a very manual system that AI techniques. If you set up a proper system, you are able to tackle it pretty easily.”
To make the process more efficient, Storoe developed an algorithm that not only looked at the images but tracked where each image was in the frame and their overall growth in the season. After working on that project, Storoe traveled with another student, Ben Paulson, to Santiago to meet with the Nilus team. It was the first time the 20-year-old had ever left the country.
“I went snowboarding in the mountains over there and that was just super fun. I’d never gone on a mountain that high, and we just tried so many different restaurants,” Storoe said. “The best day of the trip was when we actually went and visited these ice stupas up in these rural mountains where we had to drive and hike for four hours to get there.”
Now, Storoe said he’s continuing to work with Nilus as part of his senior capstone.
“My goal is to predict at scale the volume of these ice stupas,” Storoe said. “So just looking at a static camera, going from a 2D image to understanding the volume at a whole 3D realm is the goal that I’m trying to output by the end of this year.”
When asked what he was most proud of with this project, Storoe pointed to the precedent the Nilus project set for other students in the AI club.
“We showed this project to all the new AI club students this school year and showed them that, ‘Hey, if you’re a freshman, you’re a sophomore, you don’t have to limit yourself to simple projects,’” Storoe said. “You can build something incredible.”
cstein@chicagotribune.com