Pothole AI

Pothole AI is a mobile app that uses artificial intelligence to detect potholes in real time using a smartphone camera while driving. I created this project after becoming interested in how hidden infrastructure problems often go unnoticed until they become expensive, dangerous, or environmentally damaging. The app was designed around a simple idea: using AI to identify issues earlier, before they escalate into larger real-world problems. The system runs a custom-trained YOLOv11 computer vision model directly on-device using Apple Core ML, allowing detections to happen quickly without needing cloud processing or internet access. When a pothole is detected, the app automatically logs its GPS location and displays it on an interactive map for easier road monitoring and maintenance planning. Pothole AI aims to address delayed and underreported road damage, which contributes to vehicle damage, safety risks, fuel inefficiency, and unnecessary emissions from worsening infrastructure conditions. Traditional reporting systems often rely on manual submissions and delayed inspections, making it difficult to scale maintenance efforts efficiently. By turning everyday smartphones into infrastructure monitoring tools, the system offers a more accessible and scalable alternative. Key Features: • Real-time pothole detection • GPS-based mapping and logging • Fast on-device AI processing • Privacy-focused edge computing • Offline-capable monitoring • Lightweight and battery-conscious design More broadly, this project reflects my interest in engineering AI-driven systems that can detect hidden patterns early and help solve real-world problems before they become irreversible.