RoadShield AI App Development

I’m building RoadShield AI, a real-time traffic intelligence and smart-navigation system tailored for the delivery and logistics sector. The first release must run on Android phones and through a responsive web interface. My goal is to go well beyond standard congestion mapping: the engine has to spot and flag actual road-damage events as they happen, merge them with live GPS traces, crowd-sourced reports, weather feeds, and any open government traffic data, then reroute drivers instantly. Here’s what I need from you: an end-to-end solution that couples a robust AI model for detecting road damage with a slick front-end experience on both target platforms. The Android app should feel native, lightweight, and usable even on middling hardware. The web portal will act as both a dispatcher dashboard and a fallback navigation view, so it must visualise incidents, ETA shifts, and recommended alternate paths in real time. You’re free to choose the most appropriate stack—TensorFlow, PyTorch, or another ML framework on the back end; Kotlin, Flutter, React, or similar for the client side—so long as the final build delivers fast detection (<5 s from data ingest) and turn-by-turn guidance that automatically avoids confirmed damage zones. Acceptance checklist: • A trained, documented model that achieves high recall on road-damage imagery and sensor data • REST or GraphQL API serving incident data and routing suggestions • Android APK plus Play-ready bundle, fully tested across Android 9-14 • Secure, mobile-first web app hosted on a cloud instance of your choice • Source code in Git with clear build instructions and unit tests • Brief video demo proving real-time detection and rerouting in a simulated or live environment If this sounds like your kind of challenge, let’s talk timelines and milestones.

Project ID: 40423857