📄 AI Resume Extraction Engine

An AI-powered Resume Extraction Engine that extracts structured information from resume PDFs using Google Gemini AI.

The system processes uploaded resumes, extracts important details, validates the extracted data, and stores it in a SQLite database using a FastAPI backend and Streamlit frontend.

---

🚀 Features

• Resume PDF Upload
• AI-powered Resume Parsing using Gemini API
• Structured JSON Data Extraction
• PDF Text Extraction using pdfplumber
• SQLite Database Storage
• FastAPI REST APIs
• Streamlit Frontend UI
• Resume History Endpoint
• Pydantic Validation
• Error Handling & Retry Logic
• Confidence Score Support

---

🛠 Tech Stack

Backend

• FastAPI
• Python
• SQLite
• SQLAlchemy
• Pydantic

Frontend

• Streamlit

AI Integration

• Google Gemini API

PDF Processing

• pdfplumber

---

📂 Project Structure

document_ext_eng/

│── backend/
│   ├── main.py
│   ├── gemini_service.py
│   ├── extractor.py
│   ├── database.py
│   ├── models.py
│   ├── schemas.py
│   ├── utils.py
│   └── prompts.py

│── frontend/
│   └── app.py

│── uploads/
│── screenshots/
│── requirements.txt
│── README.md
│── .gitignore

---

⚙️ Installation & Setup

1️⃣ Clone Repository

git clone https://github.com/shubham99557/document_ext_eng.git
cd document_ext_eng

---

2️⃣ Create Virtual Environment

python -m venv venv

---

3️⃣ Activate Virtual Environment

Windows

venv\Scripts\activate

Linux/Mac

source venv/bin/activate

---

4️⃣ Install Dependencies

pip install -r requirements.txt

---

5️⃣ Add Gemini API Key

Create a .env file inside the backend folder.

GEMINI_API_KEY=your_api_key_here

---

▶️ Running Backend

Inside backend folder:

uvicorn main:app --reload

Backend URL:

http://127.0.0.1:8000

Swagger API Docs:

http://127.0.0.1:8000/docs

---

▶️ Running Frontend

Inside project root folder:

streamlit run frontend/app.py

Frontend URL:

http://localhost:8501

---

🔌 API Endpoints

📤 Extract Resume

POST /extract-resume

Uploads a resume PDF and extracts structured information using AI.

---

📄 Get All Resumes

GET /all-resumes

Returns all stored resumes from the SQLite database.

---

🤖 AI Workflow

1. User uploads a resume PDF
2. PDF text extracted using pdfplumber
3. Extracted text sent to Gemini API
4. Gemini returns structured JSON response
5. Response validated using Pydantic
6. Resume data stored in SQLite database
7. Extracted information displayed in Streamlit UI

---

📌 Sample JSON Output

{
  "name": "SHUBHAM RAJ",
  "email": "shubham_202300334@smit.smu.edu.in",
  "phone": "+91 6299821330",
  "skills": [
    "Python",
    "Django",
    "JavaScript"
  ],
  "education": "B.Tech CSE, SMIT",
  "projects": [
    "Kartloop",
    "Smart Parking System"
  ]
}

---

📸 Screenshots

🏠 Home Page

---

📤 Resume Upload Page

---

✅ Resume Extraction Output

---

📄 Swagger API Docs

---

📄 Swagger API Output

---

📂 All Resume API

---

📄 All Resume Output

---

🗄 Database Output

---

📁 Folder Structure

---

💻 Streamlit Frontend

---

🔮 Future Improvements

• Resume Ranking System
• ATS Score Generation
• Resume Summarization
• Multi-file Upload Support
• Skill Matching Engine
• Export to CSV/Excel
• OCR Support for Scanned PDFs

---

👨‍💻 Author

Shubham Raj
B.Tech CSE
Sikkim Manipal Institute of Technology

---