Inspiration

As students living on campus at UC San Diego, we recognized how important timely maintenance is to our everyday lives. With numerous Fix-It requests submitted by students, manually reviewing, categorizing, and prioritizing each issue can become a time-consuming process. We wanted to explore how artificial intelligence could streamline this process and help housing staff respond to maintenance problems more efficiently. This inspired us to create Fixit AI.

What it does

Fixit AI is an AI-powered maintenance request management application designed for college dormitories. When a student submits a maintenance request, Google Gemini analyzes the description, categorizes the issue, and determines its priority based on urgency and impact.

The application organizes these requests, allowing housing staff to identify pressing issues more easily and manage maintenance needs efficiently.

How we built it

We developed Fixit AI using Python and Django for our application's backend, along with HTML and CSS for the user interface. We integrated the Google Gemini API to analyze natural-language maintenance descriptions and classify requests into categories such as plumbing, electrical, HVAC, furniture, and appliances.

Through prompt engineering and structured JSON outputs, we designed Gemini to assign low, medium, or high priority levels based on the severity and impact of each issue. We also used Git and GitHub to collaborate and integrate our individual contributions.

Challenges we ran into

One of our biggest challenges was integrating different components of the application while working under the time constraints of a hackathon. As beginners, we encountered difficulties configuring Python environments, installing dependencies, connecting APIs, and merging our code through GitHub.

Another challenge involved making Gemini's classifications consistent and meaningful. We had to refine our prompts, establish clearer priority guidelines, and test different maintenance scenarios. We also had to balance our ambitious initial ideas with what we could realistically implement within the available time.

Accomplishments that we're proud of

We're proud of building an AI-powered application that addresses a practical problem affecting college students and campus housing operations.

Despite having limited experience with full-stack development, we successfully integrated Google Gemini into our project and developed a system capable of analyzing and prioritizing maintenance requests.

We're especially proud of the collaboration and problem-solving involved in bringing our individual components together into a functional prototype.

What we learned

This hackathon gave us valuable hands-on experience with Django, Python, API integration, prompt engineering, and GitHub collaboration.

Beyond the technical skills, we learned how to translate a real-world problem into a software solution. We also discovered the importance of testing, debugging, clear communication, and adapting our project goals when facing time constraints.

Most importantly, we learned that AI's value extends beyond generating responses. It can be used to automate repetitive processes and help people make more informed decisions.

What's next for Fixit AI

We hope to expand Fixit AI into a more comprehensive maintenance management platform.

Our next steps include developing an intelligent worker-assignment system that matches maintenance requests with technicians based on their expertise, availability, and workload.

We also want to implement duplicate request detection, allowing the system to recognize when multiple students report the same underlying problem, and image analysis, enabling Gemini to assess photos submitted alongside maintenance descriptions.

In the long term, we envision Fixit AI helping universities streamline maintenance operations, reduce administrative workload, and create a better living experience for students.