Dorm-fix

DormFix was inspired by the growing number of dorm maintenance requests on college campuses and the difficulty facilities teams face in efficiently prioritizing them. Student submissions are often unstructured and vary widely in clarity and urgency, requiring staff to manually review every request before taking action. As request volume increases, this process becomes time-consuming and difficult to scale.

To address this challenge, we built DormFix as an AI-assisted maintenance request system that transforms free-text student reports into structured, prioritized tickets. Students submit issues naturally through a mobile interface, while the backend uses Gemini 3 to analyze each request. The model identifies key contextual signals—such as issue type and urgency indicators—and labels requests as important or not important. In addition, Gemini generates concise AI insights explaining the priority decision and summarizing the issue for the staff member responsible for deployment.

Through this project, we learned how to integrate large language models into a production-style workflow and how critical prompt design and output consistency are for real-world usability. One of the primary challenges was handling ambiguous submissions while maintaining transparent and reliable classifications.

DormFix demonstrates how AI can enhance operational decision-making by reducing manual triage and enabling faster, more consistent maintenance responses.

Built With

• cloudinary
• express.js
• gemini
• image-picker
• jwt-auth
• mongodb
• node.js
• react-native-(expo)
• react-navigation