Designed an event-driven serverless backend using AWS Lambda, EventBridge, and DynamoDB to detect invoice fraud in real-time.

Achieved a 90% detection accuracy rate by cross-referencing invoice metadata against trusted vendor datasets using LLMs.