An open-source Codex workflow for job applications
This is an job workspace built around . It creates a candidate profile from a person's own documents, then ranks job listings by how well they match that profile. It can generate tailored resumes and cover letters for each job.
It also builds LaTeX PDF resumes and checks whether an ATS can read the text by using pdftotext. It tracks results so later s can improve. The workflow is designed to avoid making up skills, dates, employers, numbers, or education history.
It is based on the idea behind MadsLorentzen/ai-job-search, but rebuilt around s, tools, LaTeX/PDF checks, and setup instructions.
Key points
- The project is an job search assistant built for .
- It turns personal documents into a grounded candidate profile.
- It ranks job posts, then creates tailored resumes and cover letters.
- It uses LaTeX for PDF output and pdftotext to check readable text.
- It has rules to avoid inventing skills, dates, employers, metrics, or education.