// Resume optimizer. Takes a resume and a target job description, returns structured suggestions through domain-specific prompt rules and structured Claude output pipelines.
RESUMIND takes a resume and a target job description and produces structured, actionable optimization suggestions. It doesn't fabricate — it reorganizes and emphasizes what's already true.
Built with Next.js and TypeScript, the Anthropic Claude API for analysis, and MCP (Model Context Protocol) servers for development tooling. One of the earlier projects where MCP was used as a first-class architectural primitive rather than an afterthought.
The pipeline separates two distinct concerns: analysis (what does this job require?) and generation (how do I present what I already have?). This mirrors a compiler's separation of parsing from code generation — same input, different output shapes depending on the target.
Domain-specific prompt rules constrain the analysis, and structured output pipelines ensure suggestions are always parseable and renderable. The result is a tool that treats career documents the way a compiler treats source code: as text with structure, meaning, and a target representation.