CarSense

CarSense is a Flutter application that walks a driver from describing a problem to receiving a structured, AI-assisted assessment of what might be wrong with their vehicle. Version 0.1 focuses on the complete core journey rather than on hardware integrations.

Give drivers a clearer, calmer first step when something feels wrong with their car — a structured way to describe symptoms and a readable report to take to a mechanic.

Product designer and developer

• Designed the full v0.1 journey first: register, add a vehicle, start a guided diagnosis, review safety information, answer a questionnaire, receive an assessment, generate and share a report, and review history.
• Combined a structured diagnostic rule engine with an AI validation layer, so the model works inside defined rules instead of free-form guessing.
• Kept all AI calls on the backend so API keys never ship inside the mobile app.
• Architected future modules from the start but kept them clearly labelled as coming soon in the interface.

• Product definition and end-to-end user journey
• Interface design (Stitch design references)
• Flutter app architecture with Riverpod and go_router
• FastAPI backend foundation and Supabase schema migrations
• CI pipeline for Flutter analysis, tests and backend tests

• Flutter
• Dart
• Riverpod
• go_router
• FastAPI (Python)
• Supabase
• DeepSeek API
• GitHub Actions