PDF Interview Questions Generator

An end-to-end RAG-based PDF interview question generator that analyzes uploaded PDF documents and automatically generates high-quality interview questions and answers.
The system uses LLaMA 7B, Sentence-Transformers embeddings, and Pinecone vector search, and is deployed on AWS EC2 with full Docker + CI/CD automation.

---

🚀 Features

• Upload PDF documents via a web interface
• Extract and semantically analyze PDF content
• Generate interview-style questions and answers
• Retrieval-Augmented Generation (RAG) for context-aware outputs
• Export generated Q&A pairs to CSV
• Fully containerized and production-deployed on AWS

---

🧠 Architecture Overview

1. PDF Upload
   • PDFs are uploaded via a FastAPI-based web interface.
2. Text Extraction & Chunking
   • PDF content is parsed and split into semantically meaningful chunks.
3. Embedding & Vector Storage
   • Sentence-Transformers (all-MiniLM) generate embeddings.
   • Embeddings are stored in Pinecone for fast semantic retrieval.
4. Question & Answer Generation
   • Retrieved context is passed to a LLaMA 7B model.
   • The model generates structured interview questions and answers.
5. Export
   • Generated results are saved and exported as a CSV file.

---

🛠️ Tech Stack

Backend & APIs

• FastAPI
• Python
• Jinja2 (templating)

LLM & RAG

• LLaMA 7B
• Sentence-Transformers (all-MiniLM)
• Pinecone Vector Database

Infrastructure & Deployment

• Docker
• AWS EC2
• Amazon ECR
• GitHub Actions (CI/CD)

---

📂 Project Structure

.
├── .github/workflows/     # CI/CD pipelines
├── src/                   # Core application logic (RAG + LLM pipeline)
├── data/                  # Local development data
├── research/              # Experiments and exploration
├── static/                # Static assets (CSS, outputs, uploads)
├── templates/             # Jinja2 templates
├── app.py                 # FastAPI application entry point
└── README.md

---

⚙️ Running Locally

# Clone the repository
git clone https://github.com/biresh1929/PDF-Interview-Questions-Generator.git
cd PDF-Interview-Questions-Generator

# Install dependencies
pip install -r requirements.txt

# Run the application
uvicorn app:app --host 0.0.0.0 --port 8080

---

☁️ Deployment

• Containerized using Docker
• Images pushed to Amazon ECR
• Deployed on AWS EC2
• Automated build and deployment using GitHub Actions CI/CD

---

📈 Use Cases

• Interview preparation from technical PDFs
• Automated assessment content generation
• Academic and educational material analysis
• Knowledge extraction from large documents

---

🔒 Notes

• Designed for scalable semantic retrieval
• Easily extensible to support additional document formats
• Production-ready deployment setup

---

📄 License

This project is open-source and available for learning and experimentation.

---