FraudGuard AI - Ad Fraud Detection System

An intelligent ad fraud detection system that combines traditional machine learning with advanced AI analysis to identify and block fraudulent advertising traffic in real-time.

Architecture

The system uses a two-tier detection approach:

• Tier 1 (ML): Fast XGBoost-based screening for immediate decisions (<100ms)
• Tier 2 (AI): Deep analysis using Amazon Bedrock/Claude for sophisticated fraud patterns (1-3s)

Prerequisites

• AWS CLI configured with appropriate credentials
• AWS SAM CLI installed (optional, for easier deployment)
• Python 3.11
• AWS Account with permissions for:
  • API Gateway
  • Lambda
  • DynamoDB
  • S3
  • SageMaker
  • Bedrock

Project Structure

.
├── template.yaml          # AWS SAM template
├── samconfig.toml        # SAM configuration
├── docs/                 # Additional documentation
├── src/
│   ├── ingestion/        # Ingestion handler Lambda
│   ├── orchestrator/     # Fraud orchestrator Lambda
│   └── ai_analyzer/      # AI analysis Lambda
└── README.md

Setup

1. Install AWS SAM CLI (if not already installed):

   # macOS
   brew install aws-sam-cli
   
   # Or using pip
   pip install aws-sam-cli

2. Build the application:

   sam build

3. Deploy to AWS:

   sam deploy --guided

4. Get API endpoint and key: After deployment, check the CloudFormation outputs for:

   • API URL
   • API Key

Development

Local Testing

# Start local API
sam local start-api

# Invoke function locally
sam local invoke IngestionHandler --event events/event.json

Environment Variables

Set up your .env file with:

• AWS credentials
• API keys for Bedrock
• Other configuration

Documentation

Additional documentation is available in the /docs folder:

• Product Requirements: ad-fraud-detection-prd.md
• Integration Guides:
  • INTEGRATION.md
  • fraudguard-google-ads-integration.md
  • GOOGLE_ADS_IMPLEMENTATION.md
• Storage: STORAGE.md
• Testing: TESTING_BEDROCK.md
• Development: BRANCH_STRATEGY.md

Module-specific documentation:

• Dashboard: dashboard/README.md and dashboard/DEPLOYMENT.md
• Scripts: scripts/README.md
• Components: See README files in respective src/ subdirectories

Resources

• API Gateway: REST API endpoint for fraud detection
• Lambda Functions:
  • Ingestion Handler: Parse and store events
  • Fraud Orchestrator: Route to ML/AI analysis
  • AI Analyzer: Deep fraud analysis using Bedrock
• DynamoDB: Real-time event storage
• S3: Long-term data storage and training data
• SageMaker: XGBoost model endpoint
• Bedrock: Claude 3 Sonnet for AI analysis

Next Steps

1. Implement Lambda functions
2. Train and deploy XGBoost model
3. Configure Bedrock access
4. Build Streamlit dashboard
5. Set up monitoring and alerting

License

Proprietary - All rights reserved