fanmixco/career-signal

Career Signal Engine is a local app that helps job seekers review and improve their CV before applying to a specific role. The key idea is that many CVs fail not because of formatting but because they lack concrete evidence: actual numbers, measurable results, and clear descriptions of what the person did. This tool checks for that evidence before it helps you improve anything. The workflow has two stages. First, the app reads your CV (either a LinkedIn PDF export or pasted text) and gives it an Evidence Score from 0 to 100. It flags problems like vague claims with no supporting numbers, hidden career progression, tense inconsistencies, and personal details like age or citizenship that may create bias risk. You can only continue to the second stage after passing this check, or by explicitly choosing to proceed despite a low score. In the second stage, you provide a target company and job description. The app uses an AI model to build a job-specific reconstruction plan: advice on how to reframe your existing experience so it speaks to what this particular employer is looking for. It also gives a profile-match score for that company and role. You can test multiple roles against the same CV without repeating the first check. The final plan can be downloaded as a text file. The app supports three AI providers: Gemini, OpenAI, and Mistral. You choose the provider inside the app. Your API key is used for that request only and is not stored anywhere. The app runs entirely on your own computer and does not send your CV to any external database. Setup requires Node.js. You install it, download the repository, run npm install in the backend folder, and start the app with npm run dev. A Docker image is also available for a one-command start. The README walks through every step with plain instructions for Windows, macOS, and Linux.