What started as a 6th-semester college project has evolved into a professional AI showcase. Saving forests, one pixel at a time.
It was 2020. The world was in lockdown, the deadlines were tight, and I needed a 6th-semester project that wasn't just another 'To-Do List' app.
I wanted to build something with real-world impact. Forest fires were devastating ecosystems globally, and I wondered: could a simple web cam and some AI help catch them early? Armed with a dataset and a lot of determination, I dove into the world of Computer Vision.
Fast forward to today: The code is cleaner, the UI is snappier (thanks, Next.js!), and the model is smarter. But the core mission remains the same: leveraging technology to protect our planet.
The modern tools and technologies powering this project
The brain behind the operation. Powers our server-side inference engine.
The muscle. Server-side rendering, routing, and pure speed.
The safety net. Catching bugs before they catch fire.
The stylist. Making sure we look good while saving the world.
The magic. Smooth animations that make you go "ooh".
The launchpad. Deploying to the edge with a single git push.
Key metrics that showcase the project's quality (and my caffeine addiction)
Model Accuracy
98.5%
Validated on Forest Fire C4 set
Caffeine Intake
โ
Cups of coffee
Images Trained
4,823
Forest Fire C4 training corpus
Requests Served
10k+
Since v2 beta launch
The complete development journey from concept to deployment
2020
2020
The initial idea. Scouring Kaggle for datasets and training the first clunky Python model.
2023
2023
Refining the methodology. "Learning without Forgetting" became the core research focus. Published in Fire Ecology.
2025
2025
Ditching the old HTML/JS for a modern Next.js stack. Better UI, faster inference, and actual type safety.
Beyond
Beyond
Adding real-time satellite data, edge deployment support, and maybe a dark mode that is actually dark.