# ArqeloCV

> AI resume builder for software engineers.

ArqeloCV is an AI resume builder for software engineers. It converts unstructured resumes (PDF, Word DOCX, LaTeX, plain text) into structured engineering resumes, evaluates them against target job descriptions, generates evidence-grounded proposals without inventing experience, and deterministically renders clean, ATS-friendly LaTeX PDFs through a restricted Docker TeX service.

Production URL: https://arqelo.ashutoshtiwari.dev

## Core Capabilities

- **Multi-Format Extraction**: Extracts text and structure from PDF (`pdfjs-dist`), Word DOCX (`mammoth`), LaTeX (.tex), and plain text.
- **Evidence-Grounded AI Tailoring**: Tailoring suggestions link directly to verbatim evidence in the user's uploaded background. Suggestions never invent metrics, falsify timelines, or hallucinate technologies.
- **Deterministic LaTeX Rendering**: ResumeSchema is the single source of truth. LaTeX code is deterministically generated without visual drift, overlapping columns, or broken parser glyphs.
- **Restricted Docker Compilation**: Runs TeX Live in an isolated service with shell escape disabled, source limits, execution timeouts, and disposable request directories.
- **Local Privacy**: Resumes, contact details, and compiled documents are persisted in browser IndexedDB. No backend database stores candidate personal information.
- **Atomic Review Diffs**: Every change (role, summary, bullet point, skill) is isolated into an atomic diff requiring explicit user acceptance or rejection.

## Documentation & References

- [Full LLM Specification](/llms-full.txt): Comprehensive architectural, schema, and API reference for AI models and answer engines.
- [GitHub Repository](https://github.com/iashutoshtiwari/ai-resume-builder): Source code, tests, and deployment guides.