gotili/eigencv

EigenCV is a resume-building pipeline designed to stop AI tools from inventing skills or experience you do not actually have. The core idea is that your career history lives as a structured local database of verified facts, each tagged with a unique ID. When you apply to a job, an AI reads the job description and selects which entries from your database are most relevant, but it cannot write new content or change what those entries say. A Python compiler then takes the approved IDs, pulls the exact text from your database, and assembles a finished PDF resume using a LaTeX template. The pipeline includes a component the README calls a Lie Detector. If the AI tries to insert something that does not match an existing database entry, the build crashes rather than producing a resume with false information. The goal is to pass automated resume screening systems honestly, so you are not caught off guard in a technical interview by claims your resume made but you cannot back up. There are two main ways to use EigenCV. The first is through ChatGPT Plus, where you upload the project as a ZIP file, provide your old resumes, and instruct the AI to build your database and then apply to jobs by pasting in job descriptions. The second is a local developer route where you run the scripts on your own machine using an AI coding tool like Cursor, Claude Code, or Windsurf, and optionally connect a local model through Ollama so your data never leaves your computer. Supported features include multiple LaTeX resume templates, multilingual output including English and German, custom accent colors to match a target company's branding, and a cover letter generator that draws only from your verified profile. You can also reorder resume sections by editing a single variable. Setup requires Python 3.11 or higher and a LaTeX installation such as TeX Live or MiKTeX. A Docker container is available for VS Code users who prefer not to install LaTeX locally. The project is open source under the MIT license.