A Serverless Event-Driven Architecture for Intelligent Personal Finance Management with Automated Continuous Deployment on AWS Lambda

Abstract: Personal and small-business financial management increasingly demands software that is responsive, intelligent, and inexpensive to operate, yet conventional server-based applications burden developers with continuous provisioning, idle-capacity cost, and manual scaling. This paper presents a fully serverless, event-driven financial management platform that records expenses and income, enforces budgets, and delivers predictive insight while eliminating standing infrastructure. The system couples a stateless application programming interface, implemented with the FastAPI framework and adapted to a function-as-a-service runtime, to a constellation of ten independent worker functions that handle receipt optical-character recognition, categorization, budget evaluation, forecasting, notification, and audit logging in response to events. A lightweight machine-learning engine performs automatic expense categorization, regression-based forecasting, and anomaly detection, while multi-tenant isolation, token-based authentication, role-based access control, and request rate limiting protect each account. Persistence is provided by a managed NoSQL store in the cloud and a relational engine for local development, and the entire stack is released through an automated continuous-integration and continuous-deployment pipeline that executes tests, security scanning, and packaging before promotion. Experimental evaluation shows that the serverless deployment sustains a 95th-percentile latency of 640 ms at 1,000 requests per second, compared with 4,300 ms for a provisioned monolithic baseline, while warm-invocation latency averages 95 ms and the categorizer attains 88.5% accuracy. The principal contributions are a decomposed event-driven design, an embedded intelligence layer, and a reproducible automated release pipeline that together yield an elastic, cost-efficient, and maintainable financial platform.

Keywords: Serverless computing, AWS Lambda, event-driven architecture, continuous deployment, FastAPI, machine learning, expense forecasting, cloud-native applications.