Abstract
Parking at UVA is a problem that seems simple until it becomes part of a student’s daily routine. A student may leave early, still struggle to find a space, pay more than expected, or have to adjust plans because of permit rules, event restrictions, or lot closures. My technical project grew out of this everyday frustration by proposing a smarter way for students to find parking before they arrive on campus. My STS research moved in a different direction by examining AI screening systems in hiring and facial recognition. Although the two projects are not directly connected, both made me think about how software systems affect people once they become part of larger institutions. In both cases, technical design is tied to access, trust, fairness, and responsibility.
For the technical portion of my thesis, I proposed Smart UVA Parking, a mobile application designed to help students make better parking decisions around campus. The application would combine historical parking occupancy data with live crowd-sourced updates from students to predict parking availability. It would also use machine learning models such as Gradient Boosting and LSTM networks to improve these predictions over time. Beyond showing open spaces, the app would compare parking options based on cost, distance, and time limits. It would also notify students about game-day relocations, temporary lot closures, and other changes that affect parking. The goal of this project was to create a tool that reduces stress and gives students clear information in a parking system that can feel scattered and unpredictable.
In my STS research, I examined automated hiring tools and facial recognition systems through the Social Construction of Technology framework. I wanted to understand how technologies that appear objective can still reproduce bias and make accountability difficult to locate. In hiring, employer definitions of merit, success, and fit can become automated scores or rankings. In facial recognition, agency choices about databases, thresholds, and acceptable errors can become matches or alerts. My research showed that AI screening does not simply produce biased outcomes. It also spreads responsibility across vendors, employers, agencies, regulators, data practices, and users, making it difficult for affected people to understand who made the decision or who should be held responsible.
Working on these projects together helped me see engineering as more than building a system that functions correctly. Smart UVA Parking showed me that even a helpful app depends on student behavior, university policies, accurate data, and trust in shared information. My STS research showed the same lesson in a higher-stakes setting, where technical systems can influence employment, privacy, and surveillance. Responsible engineering requires asking who the system is built for, who may be left out, and who remains accountable for the results. This perspective will shape how I think about future software projects, especially when they affect people’s access to important opportunities or services.
