Students were gaming the registration system because the advising process was broken. I built a GPT-powered chatbot trained on real St. Edward's course data — a conversational alternative that actually worked.
Course registration at St. Edward's was a mess. Students didn't always know what to take, when to take it, or how it fit their degree plan. Academic advisors were overloaded, the information existed but it was buried across catalogs and outdated PDFs. Students were making uninformed decisions or delaying registration entirely.
Topper was designed to bridge that gap — not replace advisors, but handle the questions that didn't need a human in the first place.
Conversational UI designed in Figma, then built in Python/Flask. The goal: feel like asking a helpful upperclassman, not clicking through a FAQ.
Topper's chat UI — greeting flow, conversational hand-off, single-input focus.
This was a collaborative senior capstone. My partners handled the Python/Flask backend. Everything else — the research, the interface, the AI training, the mid-project infrastructure pivot — that was me.
Topper was built and tested but not publicly released — it was a senior capstone, not a production launch. But the testing validated the core idea: students could get course answers faster through conversation than through the existing system.
Topper AI was a senior capstone — explore the demo or head back to the portfolio.
