FlorianAI

Inspiration

We where inspired by link and by over work emergency dispatchers

What it does

Florian AI helps emergency responders by taking excess call that would just be put on hold and asking the caller information to properly categories their emergency. This allows dispatchers to prioritize high risk emergencies and help callers by providing verbal assistance when they would have other wise been stuck on hold.

How we built it

We used Next.js and Tailwind for the user facing side of the project and used python and flask to manage the API calls and backend.

Challenges we ran into

Learning large sets of documentation and figuring out which API features would most efficiently reach our end goal.

Accomplishments that we're proud of

The fined tuned Google Gemini model classifies emergencies based on the call transcripts in near real time.

What we learned

It's hard to tune a model and to integrate.

What's next for FlorianAI

Full call support. Meaning accept incoming phone calls and the allow responders to pick up in progress calls from the Response Dashboard

Built With

• google-genai
• hume-evi
• nextjs
• python
• tailwind