Resume Intelligence AI: Guide

resume_parser.mp4

This repository contains the Resume Intelligence AI, a sophisticated platform designed to bridge the gap between resume claims and actual technical ability. By integrating automated verification of digital footprints and AI-driven interviewing, we transform static documents into a comprehensive "Recruiter Intelligence" report.

Core AI Solution

• Resume Parsing: Accurately extracting key qualifications using NLP.
• URL Discovery & Web Crawling: Identifying relevant online portfolios and gathering real-time data from public sources.
• Code Insight Extraction: Leveraging AI to analyze repositories for tech stacks and contribution complexity.
• AI Interviewing: Contextual assessments designed to reflect true candidate abilities.

🚀 Overview

The application operates in two primary modes:

1. Resume Analyzer: Uses Mindee for OCR, PyMuPDF for link extraction, and Firecrawl for web scraping to build a rich user profile via OpenRouter LLMs.
2. ATS Scorer: Uses sentence-transformers to calculate a semantic match percentage between your profile and a job description.

---

🛠️ Setup Instructions

1. Prerequisites

Ensure you have Python 3.11+ installed. You will also need API keys for the following services:

• Mindee: For resume parsing.
• Firecrawl: For scraping external links (GitHub, LinkedIn, Portfolio).
• OpenRouter: To access AI models (e.g., arcee-ai/trinity-large-preview).

2. Installation

Clone the repository and install the required dependencies:

pip install -r requirements.txt

3. Configuration

Create a .streamlit/secrets.toml file or set environment variables for the following keys:

MINDEE_API_KEY = "your_mindee_key"
FIRECRAWL_API_KEY = "your_firecrawl_key"
OPENROUTER_API_KEY = "your_openrouter_key"

---

🧪 How to Test

Option A: Local Streamlit UI (Recommended)

Run the interactive dashboard to test the full workflow visually:

streamlit run app_ui.py

1. Tab 1 (Analyzer): Upload a PDF resume. The system will parse text, find links, scrape their content, and generate a recruiter-ready profile.
2. Tab 2 (Scorer): Paste the generated profile and a target Job Description. The system will generate a match score and a gauge chart.

Option B: Local API (FastAPI)

You can also run the backend as a standalone service:

uvicorn main:app --reload

• Health Check: GET /
• Process Resume: POST /process_resume with a JSON body {"pdf_path": "path/to/resume.pdf"}.
• Score Resume: POST /score_resume with {"user_summary": "...", "jd_summary": "..."}.

---

📂 Project Structure

• app_ui.py: Streamlit frontend implementation.
• main.py: FastAPI orchestrator for the backend.
• mindee_service.py: Logic for extracting structured data from PDFs.
• pdf_service.py: Extracts HTTP/HTTPS links from resume files.
• firecrawl_service.py: Scrapes content from portfolio or social links.
• transformer_service.py: Computes semantic similarity scores.
• openrouter_service.py: Manages multi-turn AI reasoning for profile generation.

🌐 Live Version

Test the deployed application here: Resume Intelligence AI