Fire AI

Fire AI: Landing Page Hero
Fire AI: Landing Page Features
Fire AI: Fire Map Demo

Inspiration

Wildfires in California pose a serious threat to the local community, travelers, and the environment. We wanted to create an AI-powered tool that helps people detect fire risks early and enables governments and organizations to take action as quickly as possible.

What it does

FireAI analyzes environmental conditions to predict fire risks across California. Users can click on an interactive map to check fire risk levels, helping them plan to travel safely. It also aids officials in wildfire management, alerting them to potential fire risks as early as possible.

How we built it

We built FireAI using Next.js and TypeScript for the frontend, featuring an interactive map with Leaflet.js, while the backend uses FastAPI with a Llama model for fire risk prediction, integrating real-time weather data from OpenMeteo API and mapping services from OpenStreetMap.

Challenges we ran into

We faced challenges with model integration, particularly with the Llama model's performance, and had to carefully handle data between different units and API response formats while maintaining a fast and smooth user experience.

Accomplishments that we're proud of

We successfully trained a small language model using thousands of sample responses and live data from various APIs. Our interactive fire risk map provides a straightforward way for users to assess wildfire threats and make informed decisions.

What we learned

We learned how to effectively integrate a small language model, Llama, into web applications with proper prompt engineering, and discovered the importance of error handling to delivering a user-friendly interface.

What's next for Fire AI

We plan to expand coverage beyond California, into the rest of the United States. Additionally, we aim to improve AI predictions with more advanced training and collaborate with organizations and government agencies for proactive wildfire prevention.