ShopMindAI
Automotive diagnostic assistant with ranked repair guidance and production observability.
Overview
ShopMindAI is an automotive diagnostic assistant built to help technicians and small shops move from symptom notes to structured, testable repair plans.
The platform combines VIN and OBD-code intake, retrieval context, and LLM-supported reasoning to return ranked likely causes plus confirmatory tests. It includes production health and metrics endpoints and is deployed on Azure.
Case Study
The Problem
Technicians often lose time on inconsistent first-pass troubleshooting from unstructured symptom descriptions.
Audience & Stakes
Independent mechanics and small shops where each misdiagnosis increases labor cost, parts waste, and customer downtime.
My Approach
Built a FastAPI application with structured intake, ranking logic, and deployment-safe configuration. Added health and metrics endpoints for production monitoring and deployed on Azure.
Tradeoffs
Prioritized reliability and clear response structure over a complex front-end experience to keep iteration speed high and output quality measurable.
Impact
Delivers a production-ready diagnostic workflow demo that turns free-text issues into ranked, testable repair guidance with operational monitoring built in.
Key Features
- Structured intake for VIN, OBD codes, and free-text symptoms
- Ranked diagnostic causes with confidence and confirmatory test recommendations
- FastAPI API surface with typed request/response schemas
- Operational endpoints for uptime (`/health`) and metrics (`/metrics`)
- Azure-hosted deployment for recruiter/demo access
Technical Challenges
- Balancing strict validation with flexible real-world technician input
- Combining retrieval context with model reasoning into actionable outputs
- Keeping production observability simple enough for small-team operations
Key Learnings
- Structured response design makes AI outputs far easier to operationalize
- Health and metrics endpoints are non-negotiable for production credibility
- Input validation quality has a direct effect on output reliability