Smooth Cruize

Project Description

Full-stack web application automating pothole detection using dashcam footage.

Project Overview

Engineered an AI-powered platform that automates the detection of road hazards by analyzing public-vehicle dashcam footage. The system was built to replace inefficient citizen reporting with continuous scanning, automatically generating evidence clips and precise GPS coordinates to help city maintenance teams proactively prioritize repairs.

My Role

Full-Stack Engineer (Team Project)

The Problem

Current pothole detections rely on citizen reports, which is slow and inefficient.

Key Contributions

• ✦Engineered a Python/FastAPI backend with a YOLOv8 computer vision model and OpenCV.
• ✦Automated the extraction of 3-5 second contextual evidence clips.
• ✦Developed a responsive frontend dashboard for city maintenance teams.
• ✦Leveraged existing public-vehicle dashcams to continuously scan roads, surfacing auto-generated evidence clips with key context like severity, timestamp, and GPS coordinates.
• ✦Integrated Supabase for database management to help officials efficiently track hazards and prioritize repairs.
• ✦Won 1st place at HenHacks 2026 by delivering a complete, actionable infrastructure monitoring tool.