SmartHire

AI Resume Ranking & Candidate Matching System

Overview

This project is an AI-powered resume screening system designed to automate candidate evaluation by analyzing resumes and ranking them based on their relevance to a given job description.

The system processes multiple resumes, extracts key information, computes similarity with the job description, and ranks candidates using a dynamic scoring approach.

---

Features

• Upload multiple resumes (PDF/DOCX)
• Automatic resume parsing and information extraction
• Dynamic skill extraction from resume content
• Job Description matching using TF-IDF and cosine similarity
• Candidate scoring and ranking
• Interactive dashboard using Streamlit

---

Demo video and screenshots

Setup Instructions

1. Clone the Repository

git clone https://github.com/vertexbuddy-ai-hiring/ai-resume-ranking-and-candidate-matching-system-Shriya-Guptaa.git
cd ai-resume-ranking-and-candidate-matching-system-Shriya-Guptaa

2. Install Dependencies

pip install -r requirements.txt

3. Run the Application

streamlit run app.py

---

Project Structure

project/
│
├── app.py          # Streamlit UI
├── screening.py    # Resume parsing & data extraction
├── matching.py     # Similarity computation & ranking logic
├── requirements.txt
└── README.md

---

File Description

app.py — User Interface

• Handles resume uploads and job description input
• Triggers processing pipeline
• Displays ranked candidates in a structured table

---

screening.py — Resume Screening

• Extracts text from PDF and DOCX resumes

• Identifies:

  • Email
  • Phone number
  • Years of experience

• Dynamically extracts potential skills from resume text

• Converts unstructured resumes into structured data

---

matching.py — Candidate Matching & Ranking

• Converts resume and job description text into vector representations using TF-IDF
• Computes similarity using cosine similarity
• Applies scoring logic to rank candidates
• Returns ranked results

---

System Workflow

Upload Resumes
      ↓
Screening (Parsing & Extraction)
      ↓
Matching (Similarity + Scoring)
      ↓
Ranking
      ↓
Display Results

---

Ranking Methodology

The system ranks candidates using a similarity-driven scoring approach:

1. Skill & Content Matching

• TF-IDF vectorization is used to represent both resumes and the job description
• Cosine similarity measures how closely a resume aligns with the job requirements
• This approach captures semantic relevance rather than relying only on exact keyword matches

---

2. Experience Scoring

• Years of experience are extracted using pattern matching
• Scores are normalized relative to the required experience

---

3. Final Score

• The final score is a weighted combination of similarity and experience
• Candidates are ranked in descending order of this score

---

📊 Output

The system displays ranked candidates with:

• Name
• Skills
• Experience
• Similarity Score
• Final Score
• Rank

---

Design Approach

The system is built using a modular architecture:

• Screening layer handles parsing and extraction
• Matching layer handles similarity computation and ranking

Instead of relying on predefined skill lists, the system dynamically analyzes resume content and job descriptions, making it adaptable across different domains.

---

Conclusion

This project demonstrates how NLP techniques and structured scoring can be used to automate resume screening. The system improves efficiency in candidate evaluation while maintaining a clean, modular, and scalable design.