Jobhunter Showcase

Public-facing showcase package for jobhunter, an AI-assisted job application automation system.

This package is designed to become a separate public GitHub repository without exposing private job history, personal documents, or live credentials.

Project summary

jobhunter turns fragmented job-search work into a structured workflow:

1. collect jobs from multiple sources
2. dedupe and prefilter them
3. score fit against a candidate profile
4. generate tailored German application materials
5. package the output as PDFs
6. route it into email or dashboard review workflows

Why this is interesting

• It solves a real operational problem instead of being a toy app
• It combines APIs, scraping, LLMs, file generation, and email delivery
• It shows product thinking as well as implementation work
• It balances automation with human review

Included in this package

• README.md: public landing page
• CASE_STUDY.md: deeper product and engineering write-up
• jobhunter.py: Python backend snapshot from the working project
• pyproject.toml: Python project metadata for the backend
• dashboard/: sanitized Next.js dashboard demo
• .env.example: placeholder environment variables with explanations
• data/: empty JSON stores that mirror the local app shape
• assets/Sample_Candidate_CV.pdf: fake example CV for demo purposes
• PUBLISH.md: steps for turning this folder into a separate public repo

Intentionally excluded

• real jobs.json
• real crawl_history.json
• real stepstone_raw.json
• real generated letters and PDFs
• .env
• .gmail_token.json
• personal contact details

Suggested GitHub repo metadata

Repo names:

• jobhunter-showcase
• ai-job-application-automation
• jobhunter-case-study

Suggested description:

AI-assisted job application automation with multi-source ingestion, fit scoring, tailored German cover letters, PDF packaging, and email workflows.

Suggested tagline:

A local-first workflow for discovering jobs, screening fit, generating application material, and packaging everything for delivery.

What to show in a public repo

• the problem the system solves
• the architecture and workflow
• the model split and why it exists
• screenshots or demo assets later
• only fake sample data

Good next additions

• dashboard screenshots
• a short demo GIF or video
• an architecture diagram
• a small runnable demo mode with fake data