AI Resume Mining & Candidate Matching System

An end-to-end AI-powered recruitment platform that automatically processes resumes, extracts candidate information, generates semantic embeddings, performs intelligent candidate-job matching, and provides analytics for recruiters.

Overview

Recruitment teams often spend significant time manually reviewing resumes and matching candidates to job requirements.

This project leverages Natural Language Processing (NLP), Embedding Models, Vector Search, and Machine Learning techniques to automate candidate screening and recommendation.

The system can:

• Parse Resume PDFs
• Extract Skills, Education, and Experience
• Generate Semantic Embeddings
• Store Candidate Vectors using FAISS
• Match Candidates to Job Descriptions
• Rank Candidates by Similarity Score
• Identify Missing Skills (Skill Gap Analysis)
• Visualize Recruitment Insights

---

Features

Resume Processing

• PDF Resume Parsing
• Text Extraction
• Resume Cleaning & Normalization
• Structured Candidate Profile Generation

Information Extraction

• Skill Extraction
• Education Extraction
• Experience Extraction
• Keyword Extraction

AI Matching Engine

• Sentence Transformer Embeddings
• Semantic Similarity Search
• Candidate Ranking
• Job-Candidate Matching

Analytics

• Candidate Statistics
• Skill Distribution Analysis
• Recruitment Reports
• Interactive Dashboards

API Services

• Resume Upload API
• Candidate Search API
• Job Matching API
• Recommendation API

---

System Architecture

Resume PDF
    │
    ▼
Resume Parser
    │
    ▼
Text Cleaning
    │
    ▼
Information Extraction
(Skills, Education, Experience)
    │
    ▼
Sentence Transformer
    │
    ▼
Embedding Vector
    │
 ┌──┴─────┐
 │        │
 ▼        ▼
MongoDB  FAISS
 │        │
 └──┬─────┘
    ▼
Matching Engine
    ▼
Ranking Engine
    ▼
Analytics Dashboard

---

Project Structure

AI_ResumeMining_CandidateMatchingSystem
│
├── app
│   ├── analytics
│   ├── api
│   ├── core
│   ├── database
│   ├── embeddings
│   ├── extraction
│   ├── matching
│   ├── resume_processing
│   ├── models
│   ├── schemas
│   ├── services
│   ├── utils
│   └── vector_store
│
├── data
├── faiss_index
├── docker
├── logs
├── notebooks
├── tests
│
├── requirements.txt
├── README.md
└── run.py

---

Tech Stack

Backend

• Python
• FastAPI
• Uvicorn

Database

• MongoDB
• PyMongo

NLP & AI

• PyTorch
• Transformers
• Sentence Transformers

Vector Search

• FAISS

Data Processing

• Pandas
• NumPy
• Scikit-Learn

Visualization

• Streamlit
• Plotly

Utilities

• Loguru
• Python Dotenv

---

Installation

Clone Repository

git clone https://github.com/your-username/AI-Resume-Mining-System.git

cd AI-Resume-Mining-System

Create Environment

conda create -n ai_resume python=3.11

conda activate ai_resume

Install Dependencies

pip install -r requirements.txt

---

Environment Variables

Create a .env file:

MONGO_URI=mongodb://localhost:27017

DB_NAME=resume_mining_db

MODEL_NAME=sentence-transformers/all-MiniLM-L6-v2

FAISS_INDEX_PATH=faiss_index/candidate.index

---

Running the Project

Start FastAPI Server

uvicorn app.main:app --reload

Start Dashboard

streamlit run app/analytics/dashboards/dashboard.py

---

API Endpoints

Resume APIs

Method	Endpoint
POST	/api/v1/resumes/upload
GET	/api/v1/resumes
GET	/api/v1/resumes/{id}

Matching APIs

Method	Endpoint
POST	/api/v1/match
GET	/api/v1/rankings
GET	/api/v1/recommendations

---

Machine Learning Pipeline

Resume Processing

1. Upload Resume
2. Extract Text
3. Clean Text
4. Extract Information

Embedding Pipeline

1. Candidate Profile
2. Sentence Transformer
3. Embedding Generation
4. FAISS Storage

Matching Pipeline

1. Job Description Embedding
2. Candidate Embedding
3. Cosine Similarity
4. Ranking
5. Recommendation

---

Future Improvements

• RAG-based Candidate Search
• LLM-powered Candidate Summarization
• Multi-language Resume Support
• Real-time Candidate Recommendations
• Cloud Deployment
• Kubernetes Support
• AWS/GCP Integration

---

Author

Nguyen Hoang Khang

Information Technology Student

Interested in:

• Data Engineering
• Artificial Intelligence
• Machine Learning
• NLP
• Backend Development

---

License

This project is intended for educational and portfolio purposes.