Full-stack application that analyzes resumes against job descriptions and generates tailored, ATS-friendly PDFs. It combines a React interface with a FastAPI service, document parsing, structured matching, factuality validation, and Azure OpenAI-powered tailoring.
This project is structured as a decoupled application with a FastAPI backend and a Vite-powered React frontend. The backend separates extraction, matching, tailoring, document rendering, and validation into focused micro-services.
Key Features
Intelligent Extraction: Uses Azure OpenAI and PyMuPDF to convert unstructured PDF resumes into typed data models.
Contextual Analysis: Compares candidate profiles with role requirements to output explainable match scores and identify skill gaps.
Automated Tailoring: Dynamically rewrites the resume to highlight relevant experience without inventing facts, generating ATS-friendly PDF or Markdown files.
Factuality Validation: Built-in validation checks before the generated output is returned to prevent AI hallucinations.
AI-Assisted Workflow: Developed by actively using AI coding assistants to accelerate boilerplate generation and debugging, ensuring full ownership and validation of the generated architecture.