A group of UVA students from the McIntire School of Commerce recently demonstrated their innovative prowess at gHacks AI Hackathon event held at Google’s offices in Washington, DC, as part of AI Week. Alumna Virginia Sharma (McIntire ’99), Head of Brand and Product Marketing for Google Public Sector, connected with McIntire Dean Nicole Thorne Jenkins, resulting in an invitation to Professor Brent Kitchens and his IT students, along with Professor Chip Ransler and his students. The event provided an opportunity to tackle pressing issues related to addressing wildfires and floods across Virginia, leveraging AI to propose impactful solutions.
Sharma praised the students for their diverse and holistic approaches. “The delegation of students from the McIntire School of Commerce fully immersed themselves into solving a key issue many of them faced firsthand in the state: wildfires,” she says, acknowledging the team members’ ability to tackle the issue from various stakeholder lenses, with some focusing on emergency workers, others on resident resources. “And some went a step further to focus on the most vulnerable communities. They were partnered with a Google engineer mentor to validate their solution design, and the mentors absolutely embraced their teams and offered extra time to ensure all their questions were answered.”
One challenge the students faced was the open-ended nature of the hackathon. Caitlin Biggs (A&S ’22, M.S. in Commerce ’24), who worked with a team of fellow M.S. in Commerce students, highlighted the difficulty in creating solutions without clear guidelines. “Working without any framework made it difficult to assess what it was they were looking for,” she says.
Perry Gibbs (McIntire ’24), a member of one of the winning teams, echoed this sentiment, noting the ambiguity and the steep learning curve of using AI for natural disaster contexts. “The hackathon was challenging because it pushed my team to develop a novel solution to a highly ambiguous problem using tools that we are relatively inexperienced with,” he says. “We’ve really only been learning to work and learn with AI for a few years and have never been asked to utilize it within the context of a natural disaster.”
Innovative Ideas
Despite these challenges, the students developed impressive proposals. Biggs’ team proposed a three-stage approach to relay critical information during wildfires, including automated dispatch alerts, smart evacuation routes, and relief area coordination. “We used AI and data from IoT devices, satellites, and social media to detect wildfires early and send timely alerts,” Biggs explains. “Predictive machine learning algorithms were key to ongoing risk assessment and resource optimization.”
Gibbs’ team focused on using AI-enabled drones to predict wildfire paths, a solution inspired by interviews with local Charlottesville firefighters, who helped the team narrow its focus. “Spread prediction functionality was a common desire across interviews. We aimed to empower firefighters to take preventive action,” he says, adding how rewarding it was to have the chance to work on a meaningful problem with a diverse group of students.
Manish Dahal (McIntire ’24) and Anneloes Knol (McIntire ’24) both emphasized the steep learning curve and the importance of understanding Virginia’s wildfire and flood issues. “We read articles and consulted friends in emergency services to gain insights,” Dahal said. Knol’s team proposed a wildfire alert and safety platform using GPS and a GenAI-powered chatbot to enhance personal safety during wildfires. “It was enjoyable to see how AI can better the world,” Dahal adds.
Viva Naini (McIntire ’24) underscored the challenge of building upon existing solutions. His team proposed the creation of software they dubbed “Skyfire AI” that would use data such as topological information, wind speed, and other relevant factors to predict where wildfires would spread, utilizing Google Vertex AI. “Skyfire AI would then interface with drones that fire stations would have on site to scout fires and simultaneously provide data to refine the model using multi-object tracking.”
For Knol, relying on GPS information was key; the system would help to accurately locate users and send instant notifications when people enter wildfire-prone areas. “The notification and link to the platform would ensure that individuals in high-risk zones receive vital information quickly, empowering them to make well-informed decisions in emergencies,” she says, explaining that the platform’s advanced GenAI-powered chatbot would provide real-time advice and guidance. “Using Google Gemini multi-modal technology, the chatbot could respond to text or analyze images and videos that people would upload. This would help people handle a wildfire emergency adequately and help prevent the number of wildfires. Humans cause 80-90% of wildfires, and being aware of being in a high-risk area and having real-time and personal advice can help mitigate the risks of causing a wildfire due to unsafe practices.”
A Valuable Experience
Beyond the technical challenges, the students relished the experience of working with experts at Google and attending the AI Talks conference. “It was an amazing experience to be surrounded by members of the U.S. government and major corporations speaking on such an important topic,” Biggs says.
Gibbs, who calls himself “someone who likes to spend a lot of time outdoors,” and as such, is greatly concerned with wildfires, was thrilled to be able to take part in the hackathon and “use everything I’ve learned at McIntire to hopefully reduce future wildfire spread.” He also highlighted how courses at McIntire were instrumental in the success of her contributions: “Business Analytics with Python taught by Professor Jingjing Li and Entrepreneurship with Professor Ransler were the two classes that helped me most in the hackathon. Professor Li’s class helped me understand potential use cases and limitations of AI models, and Professor Ransler’s class helped me understand how to develop and pitch entrepreneurial ideas.”
Overall, the hackathon provided McIntire students a unique platform to apply their skills, collaborate with peers, and receive mentorship from industry leaders. As Sharma expressed, the inventiveness of the students offered her a particularly gratifying moment. “As an alum, it was a proud day for me to host some of the brightest students in the country at the Google office and have them impress my colleagues, who are some of the brightest minds in the industry. Special thanks to Dean Jenkins for connecting us to the opportunity and Professor Ransler for organizing the group! I hope we can make this an ongoing engagement between Google Public Sector and the University,” she says, reflecting on the success of the event and its promise for future collaborations.
