AI Resume Tailoring

Instead of overhauling resume customization all at once, it's structured around field-by-field mapping and change review.

Users upload their resume and paste in a JD, and the app suggests parts to adjust for each item. Each suggestion comes with why it's being changed, and users can directly approve/reject/edit it.

The core flow is as follows.

• Generate a polished .docx resume
• Quality checks for role fit, evidence, clarity, ATS readiness
• local model support via Ollama
• no blind one-shot rewriting
• Preserve metrics, impact, evidence

The stack consists of React, Express, Ollama, the local model qwen3:8b, and .docx generation/editing. It's closer to an experiment aimed at reducing the exaggeration or blurring typical of LLMs, while still keeping the resume in an actual deployable format.