🎓 Albab AI-Powered Exam System

A production-ready, scalable exam management engine powered by Groq's Llama 3 for automated evaluation of MCQs, short answers, and complex essay questions.

---

📖 Overview

The Albab AI-Powered Exam System is a sophisticated platform designed to streamline the assessment process. By leveraging State-of-the-Art LLMs via Groq's high-speed API, it provides near-instantaneous, objective grading for both structured and unstructured responses.

🌟 Key Highlights

• 💨 Blazing Fast: Powered by Groq for sub-second AI responses.
• 🛡️ Resilient: Built-in crash recovery and background evaluation queues.
• 📊 Admin Control: Full dashboard for exam monitoring and analytics.
• 🐳 DevOps Ready: Includes specialized Docker configurations and CI/CD pipelines.

---

✨ Features

🎯 Intelligent Evaluation

• AI-Powered Grading: Automated scoring using Groq's Llama 3 70B model.
• Hybrid Question Support:
  • MCQs: Single and Multi-select support with partial grading.
  • Subjective: Short answers and Essays with contextual feedback.
• Feedback Engine: Provides candidates with constructive, AI-generated reasoning for their scores.

🔒 Enterprise Security

• Anti-Cheat Measures: Server-side time enforcement and session locking.
• Resume Capability: Candidates can resume exams after network failures or browser crashes.
• Data Integrity: SQLite with Write-Ahead Logging (WAL) ensures zero data loss during concurrent access.

🛠️ Admin Capabilities

• Flexible Exam Design: Create exams via AI generation or manual input.
• Multi-Lingual: Native support for English and Bengali assessments.
• Live Monitoring: Track active sessions, time remaining, and submission status in real-time.
• Negative Marking: Granular control over penalty scoring per section.

---

🏗️ System Architecture

graph TD
    User((Candidate)) -->|FastAPI| WebServer[App Server]
    Admin((Admin)) -->|Dashboard| WebServer
    WebServer -->|Requests| Queue[Background Priority Queue]
    Queue -->|Process| Groq[Groq AI API]
    WebServer <-->|Persistence| DB[(SQLite WAL)]
    Queue <-->|State| DB

Loading

📁 Technical Blueprint

1. Core (app.py): FastAPI engine handling high-concurrency requests.
2. Logic (groq_analyzer.py): Intelligent prompt engineering for grading and generation.
3. Concurrency (evaluation_queue.py): Custom-built priority queue with exponential backoff and retry logic.
4. Data (db.py): Optimized thread-safe database layer.

---

🚀 Getting Started

📦 Quick Run (Docker Compose)

The fastest way to get started in a production-like environment:

docker-compose up -d

Access the app at: http://localhost:7894

🐍 Local Development Setup

1. Environment Configuration

Clone the repo and initialize your environment:

git clone https://github.com/rashedulalbab253/GenAI-Assessment-Engine.git
cd ai-exam-system
python -m venv venv
source venv/bin/activate  # Or `venv\Scripts\activate` on Windows
pip install -r requirements.txt

2. Environment Variables

Create a .env file from the template:

cp .env.example .env

Update these critical values:

• API_KEY: Your Groq API key.
• ADMIN_SECRET_KEY: A secure password for the dashboard.
• PORT: Default is 8000.

3. Launch

python main.py

---

🌍 CI/CD & Deployment

Automated Build Pipeline

This repository is configured with GitHub Actions (.github/workflows/docker-publish.yml) to automatically build and push Docker images to Docker Hub.

Setup Requirements:

Ensure your GitHub Repository has the following secrets:

• DOCKER_USERNAME: Your Docker Hub handle.
• DOCKER_PASSWORD: Your Docker Hub Access Token.

---

📊 Usage Guide

Admin Workflow

1. Login: Access /admin/login using your ADMIN_SECRET_KEY.
2. Design: Use the "Create Exam" tool. Define sections (e.g., "General Knowledge", "Coding").
3. Deploy: Copy the generated link and share it with candidates.
4. Monitor: Keep the dashboard open to watch candidates take the exam in real-time.

Candidate Workflow

1. Entry: Join via the unique exam link.
2. Assessment: Complete questions; answers are auto-saved to prevent loss.
3. Completion: Submit the exam. View status at the evaluation tracker.

---

🛡️ Security Best Practices

• Secrets: Never commit your .env file.
• API Keys: Use Groq personal access tokens with limited scope.
• Database: Periodically back up the exam_system.db file.
• Production: In production, set RELOAD=false in your .env.

---

🤝 Contributing

We welcome contributions!

1. Fork the project.
2. Clone your fork.
3. Create a Feature Branch.
4. Submit a Pull Request.

---

📄 License

Distributed under the MIT License. See LICENSE for more information.

---

Developed with precision by Rashedul Albab 🚀