Workshop (Vehicle Maintenance) - Fast, Fluid, & Functional: Guided Build Night w/ LiquidMetal AI
AI Tinkerers - Seattle
Hackathon Showcase

Workshop (Vehicle Maintenance)

Team led by an Amazon Engineering Manager (Michael Holst) with 26 years' experience in Java/TypeScript/Kotlin, AWS (EC2/Fargate, DynamoDB), OpenSearch, Kafka, and LLM-driven dev tooling.

1 member

Project: A web application for tracking maintenance, parts, and shops for cars, motorcycles, boats, and other DIY/hobby vehicles.

Creative Aspect:

I have not found a dead simple tool to handled this for hobby (ie. “money pit”) vehicles, that owners tend to customize and handle maintenance on their own.

Methodology:

I have been using this problem statement as my go-to evaluation of AI-first coding tools, as it includes a real-world case (I myself need it) with nuances of target audience and usage (ie. should be general but geared for DIY or hobby vehicles mostly).

I mostly one shot the description to evaluate how the tool prompts with the nuances and designs the relational entities, for example how it detects the finer grained parts and maintenance relationship to the vehicles.

Stack:

The stack chosen is fairly standard and was chosen for ease of deploying to platforms like Vercel and Supabase for reliability and scaling:

  • Frontend: Next.js 14 with TypeScript
  • Styling: Tailwind CSS
  • Database: SQLite with Prisma ORM
  • Icons: Heroicons

PC dev env setup with WSL and Claude Code, framework evaluation

Claude Code helped diagnose that the Raindrop-Code brew binary was MacOS and would not run on Linux x86-64 (that alone probably saved hours) LiquidMetal AI