If you鈥檝e ever entered the parking garage behind the USF Library on a busy morning wondering if there鈥檚 an empty space, you know it鈥檚 a gamble that might make you late for class.
Heather Ho, a senior in USF鈥檚 Bellini College of Artificial Intelligence, Cybersecurity and Computing, and the Operational Technology Access and Integrations team developed an app for that.
Ho spent hours watching recorded video footage of cars entering and exiting USF parking garages. It was how she trained an AI system that identifies cars entering and exiting parking garages to estimate available parking spaces.
In doing so, she鈥檚 helping solve one of campus life鈥檚 biggest frustrations. And she鈥檚 applying her classroom learning to a real-world problem while gaining invaluable tech experience.
Not bad for someone who once described herself as 鈥渟hy,鈥� 鈥渓ost,鈥� and unsure of her direction.
High school, interrupted
Ho, like many students at the time, looked forward to her high school鈥檚 spring break in 2020. Then came COVID-19. What was supposed to be a week of vacation turned into weeks of school shutdowns.
She never returned to that campus as a student. Instead, Ho finished high school online and started her USF classes in 2020 as a commuter student who travelled from the Town 鈥楴 Country area of Tampa.
鈥淚 had to drive around 35 to 40 minutes every day to get to class,鈥� Ho said. 鈥淚 would just come to campus, go to class and leave. I didn鈥檛 really stay or talk to anyone. I was kind of lost. I was very shy. I didn鈥檛 talk to anyone. I was often confused.鈥�
She attended classes but did little else, unsure how to navigate college as the first in her family to go. She didn鈥檛 realize what she was missing.
moving away from a transactional experience
As Ho moved deeper into her coursework, she started finding classes that challenged her thinking and professors who pushed her to grow. A turning point came in a database class.
鈥淚 kind of struggled with that class a little bit,鈥� she said. 鈥淏ut I realized that I enjoyed data and databases. So, I decided that this is something that I want to do in the future.鈥�
She found her rhythm academically. Courses in programming, data and web systems helped Ho understand what aspects of computing she enjoyed most and gave her direction at a time when she was unsure of her path.
Once unsure of where she fit, she slowly began taking ownership of her experience. She began asking questions. She talked to other students. She started to find her fit.
Finding community
As Ho became more comfortable, she gained friendships and found support. She joined the Vietnamese Student Association and took part in its activities and programs. And got involved in different Asian organizations at USF as well.
She met friends. She joined study groups. She attended events.
鈥淪ophomore year was one of my craziest years. I was involved in so many things and so USF became my second home. I was always here.鈥�
Ho wasn鈥檛 looking for a job at the time, but an opportunity presented itself thanks to these organizations.
Getting an IT job on campus
鈥淭he president of VSA鈥� he actually had a position at USF IT before, and he was like, 鈥楬ey, there鈥檚 a job position opening at USF IT and if you want, I can refer you,鈥欌�� he said.
She applied, got the gig, and it gave her the first real experience with AI and computing outside of class.
One of her first tasks: help her team build and refine a machine-learning system to estimate the volume of open parking spaces in high-traffic garages: the Crescent Hill Garage, the Collins Boulevard Parking Facility and the Richard A. Beard Parking Garage. Though it is still in Beta testing, . It shows student permit availability in these facilities and could help fellow commuter students avoid one of their greatest frustrations.
The system had to be trained. It needed to know the difference between a golf cart travelling through a garage, a person and other things that it could mistake for a car. To develop the tracker, Ho and her IT team reviewed tens of thousands of parking-lot video frames, tagged cars and checked predictions to make sure the system was accurate.
The team also had to help the camera system to recognize that cars overlapping one another 鈥� like one entering and one exiting a garage 鈥� were not a single vehicle.
The work was tedious at times, but she understood the purpose behind it. The more data that is entered, the more AI recognizes and learns. The more it learns, the more accurate its assessment on availability.
鈥溾�淚 learned that labeling takes a long time,鈥� she said. 鈥淏ut it helps improve the model. My main goal was to train the machine,鈥� she said. The impact was immediate, accuracy jumped from 鈥�50 to 60 percent to 90 to 95 percent.
A clearer path forward
She was contributing to a project that thousands of students rely on every day. And like the AI system she helped train, Ho learned to sort through uncertainty, recognize signals that mattered and find her way forward. She's shaping her path deliberately now, not waiting for opportunities, but seeking them.
鈥淟ooking back, I think, wow, I would be so different now if I didn't put myself out there.鈥�
The parking app she helped train may help students find open spaces but for Ho, it helped her find something just as meaningful: her path. She hopes to gain even more real-world experience through an off-campus internship where she can continue building the skills that excite her most: working with code, data and AI tools that make people鈥檚 lives easier.