- Project Overview
- 🌟 Features
- 🔗 Live Demo
- 🖥️ Frontend
- ⚙️ Backend
- ☁️ Infrastructure as Code
- 🛠️ Technology Stack
- 🚀 Deployment
- 💰 Cost Analysis
- 🔒 Security Features
- 📊 Performance Metrics
- 🎯 Future Improvements
QuizCraft is a cloud-native, AI-powered quiz generation platform that improve the way students and educators create personalized learning assessments. Built with modern serverless architecture on AWS, it automatically generates high-quality multiple-choice quizzes from uploaded PDF documents or custom topics using OpenAI's API.
To provide students and educators with an intelligent, scalable solution for creating personalized quizzes that adapt to specific learning content, saving time and enhancing educational outcomes.
- 🤖 AI-Powered Quiz Generation using OpenAI API for intelligent question creation
- 📄 PDF Processing with automatic text extraction and content analysis
- ⚡ Serverless Architecture ensuring scalability and cost-effectiveness
- 🔐 Enterprise-grade Security with AWS Cognito authentication and KMS encryption
- 📊 Real-time Analytics with performance tracking and insights
- 🌍 Global Content Delivery through CloudFront CDN
- 💼 Infrastructure as Code with complete Terraform automation
- Smart Quiz Generation: Upload PDFs or describe topics to generate personalized MCQ quizzes
- Instant Feedback: Get immediate results with detailed explanations for correct answers
- Performance Tracking: Monitor quiz attempts, scores, and improvement over time
- Flexible Learning: Practice with AI-generated questions tailored to your study material
- Email Notifications: Receive alerts when quizzes are ready and completion confirmations
- Content-Based Quizzes: Generate assessments directly from teaching materials
- Analytics Dashboard: Track student performance with QuickSight visualizations
- Bulk Quiz Creation: Process multiple PDFs for comprehensive course assessments
- Customizable Difficulty: Adjust quiz complexity based on student levels
- Frontend: Deployed on AWS CloudFront + S3
- Backend API: Serverless Lambda functions via API Gateway
- Database: Amazon DynamoDB with global secondary indexes
The frontend is developed using React 18 with modern hooks and functional components, implementing responsive design patterns and AWS Amplify integration.
- User Authentication: Seamless login/signup with AWS Cognito
- File Upload Interface: Drag-and-drop PDF upload with progress indicators
- Interactive Quiz Taking: Real-time quiz interface with timer and progress tracking
- Results Dashboard: Detailed score breakdowns and answer explanations
- Responsive Design: Optimized for desktop, tablet, and mobile devices
Navigate to frontend directory cd frontend/quizcraft-frontend
Install dependencies npm install
Configure environment variables Create .env file with: REACT_APP_API_ENDPOINT=your_api_gateway_endpoint REACT_APP_AWS_REGION=us-east-1
Start development server npm start
Application will be available at http://localhost:3000
- React 18 with hooks and functional components
- Material-UI for consistent design system
- AWS Amplify for authentication integration
- React Router for client-side navigation
- Axios for API communication
The backend is built using serverless AWS Lambda functions, providing automatic scaling and cost-effective compute resources.
- Validates user input and uploads PDFs to S3
- Sends processing tasks to SQS queue
- Returns immediate response to user
- Processes PDFs using PyPDF2
- Calls OpenAI API for question generation
- Stores quiz data in DynamoDB
- Sends completion notifications via SNS
- Retrieves user's quiz history
- Supports pagination and filtering
- Returns formatted quiz data
- Evaluates user answers
- Calculates scores and feedback
- Stores attempt data in DynamoDB
- Aggregates user performance metrics
- Provides analytics data for dashboard
Navigate to backend directory cd backend/
Package Lambda functions Each function has its own deployment package.
Environment Variables (configured via Terraform): OPENAI_API_KEY=stored_in_aws_secrets_manager S3_BUCKET=pdf_storage_bucket QUIZZES_TABLE=dynamodb_quizzes_table ATTEMPTS_TABLE=dynamodb_attempts_table SNS_TOPIC_ARN=notification_topic
- Partition Key: quiz_id (String)
- Attributes: user_id, quiz_content, topic, created_at, attempts_count
- GSI: UserIdIndex for user-specific queries
- Partition Key: attempt_id (String)
- Attributes: quiz_id, user_id, answers, score, completed_at
- Partition Key: user_id (String)
- Sort Key: topic_id (String)
- Attributes: topic_name, source_identifier, created_at
All AWS resources are defined and managed using Terraform, ensuring reproducible and version-controlled infrastructure.
Frontend Hosting S3 Bucket for static assets
CloudFront distribution for global delivery
Origin Access Identity for security
Authentication & Authorization Cognito User Pool with email verification
JWT token-based authentication
IAM roles with least-privilege access
Compute & Processing 8 Lambda functions for business logic
SQS queue for asynchronous processing
API Gateway with Cognito authorizer
Data Storage DynamoDB tables with GSIs
S3 bucket for PDF storage with lifecycle policies
KMS encryption for data at rest
Monitoring & Notifications CloudWatch logs and custom metrics
SNS topic for email notifications
CloudWatch alarms for error monitoring
Navigate to terraform directory cd terraform/
Initialize Terraform terraform init
Review planned changes terraform plan
Deploy infrastructure terraform apply -auto-approve
Deploy frontend (after initial infrastructure) ./deploy.sh
- Runtime: Python 3.9 with Boto3 SDK
- AI Integration: OpenAI API for question generation
- PDF Processing: PyPDF2 for text extraction
- Authentication: JWT tokens via AWS Cognito
- Database: DynamoDB with on-demand scaling
- Queue: Amazon SQS for asynchronous processing
- Notifications: Amazon SNS for email alerts
- Framework: React 18 with functional components
- UI Library: Material-UI for responsive design
- Authentication: AWS Amplify Auth
- HTTP Client: Axios with interceptors
- Routing: React Router v6
- Build Tool: Create React App with Webpack
- Compute: AWS Lambda (serverless)
- Storage: Amazon S3 with lifecycle policies
- Database: Amazon DynamoDB (NoSQL)
- CDN: Amazon CloudFront
- API: Amazon API Gateway
- Monitoring: Amazon CloudWatch
- Analytics: Amazon QuickSight
- Security: AWS KMS, IAM, Cognito
- IaC: Terraform for complete infrastructure automation
- Version Control: Git with modular architecture
- Monitoring: CloudWatch Logs, Metrics, and Alarms
- Security: Encryption at rest and in transit
- AWS CLI configured with appropriate permissions
- Terraform >= 1.0 installed
- Node.js >= 18 for frontend development
- OpenAI API key for quiz generation
-
Clone Repository git clone cd quizcraft
-
Deploy Infrastructure cd terraform terraform init terraform apply -auto-approve
-
Build and Deploy Frontend cd ../frontend/quizcraft-frontend npm install npm run build
-
Sync Frontend to S3 cd ../../terraform ./deploy.sh
The deployment script automatically:
- Builds the React application with API endpoints
- Uploads assets to S3
- Invalidates CloudFront cache
- Provides live URLs
| Service | Estimated Cost |
|---|---|
| AWS Lambda | $2.50 (2M invocations) |
| Amazon S3 | $0.60 (storage + requests) |
| Amazon DynamoDB | $3.00 (reads/writes + storage) |
| API Gateway | $1.00 (3K API calls) |
| CloudFront | $0.30 (frontend traffic) |
| CloudWatch | $1.50 (logs + metrics) |
| Amazon SNS | $0.10 (email notifications) |
| Total | ≈ $9.00/month |
- Serverless-first: Pay only for actual usage, no idle compute
- S3 Lifecycle Policies: Auto-archive old PDFs to Glacier after 30 days
- DynamoDB On-Demand: Scales automatically with traffic patterns
- Minimal Monitoring: Strategic CloudWatch alarms to avoid overcharges
- ✅ AWS Cognito Authentication with email verification
- ✅ JWT Token-based Authorization for API access
- ✅ IAM Least-Privilege Policies for resource access
- ✅ KMS Encryption for data at rest (S3, DynamoDB)
- ✅ HTTPS/TLS for all data in transit
- ✅ Input Validation in Lambda functions
- ✅ VPC Endpoints for private database access
- ✅ CloudFront Security Headers for frontend protection
- All PDF uploads encrypted in S3 using KMS
- DynamoDB tables encrypted with customer-managed keys
- Secrets Manager for secure API key storage
- Regular security monitoring with CloudWatch
| Metric | Target | Achievement |
|---|---|---|
| API Response Time | < 200ms | ✅ 120ms average |
| Quiz Generation | < 20 seconds | ✅ 18 seconds average |
| System Availability | 99.9% | ✅ 99.95% |
| Lambda Cold Start | < 1 second | ✅ 800ms average |
- CloudWatch Dashboards for real-time metrics
- Custom Alarms for error rates and latency
- Log Aggregation for debugging and analysis
- Performance Tracking for continuous optimization
- ⏱️ Detailed Time Tracking: Monitor time spent per question for learning analytics
- 🎨 Visual Content Support: Add images and diagrams to quiz questions
- 🎯 Adaptive Difficulty: AI-powered difficulty adjustment based on user performance
- 👥 Multi-tenant Support: Role-based access for educational institutions
- 📱 Mobile Application: Native iOS/Android apps with offline capability
- 🔄 Question Bank: Build reusable question libraries from past quizzes
- 📈 Advanced Analytics: Machine learning insights for learning pattern analysis
- Container Migration: Consider ECS Fargate for better cold-start performance
- Multi-Region Deployment: Global availability with disaster recovery
- Advanced Caching: Redis integration for improved response times
- Batch Processing: Parallel PDF processing for large document uploads
- ✨ 100% Serverless Architecture with automatic scaling
- 🚀 Sub-60 Second Quiz Generation from PDF content
- 🔒 Enterprise Security Standards with comprehensive encryption
- 💰 Cost-Effective Design under $10/month for 100+ users
- 📊 Real-time Monitoring with proactive alerting
- 🌍 Global Content Delivery via CloudFront CDN
- 🤖 AI Integration with intelligent question generation
Jaykumar S. Prajapati
Computer Science Student at Dalhousie University
- Course: CSCI 5411 - Advanced Cloud Architecting
- Email: jy952466@dal.ca
- LinkedIn: Linkedin
This project is developed for academic purposes as part of CSCI 5411 coursework at Dalhousie University. All rights reserved.
- Dalhousie University for providing the learning platform
- AWS for comprehensive cloud services
- OpenAI for AI-powered question generation
- Open Source Community for excellent tools and libraries