Arquiteturas distribuídas e resiliência de dados: o uso de big data e machine learning na detecção de anomalias em transações financeiras

This article analyzes the evolution of software architecture within the financial context, focusing on the transition from monolithic systems to distributed microservices and their impact on transactional security. The research investigates how the integration of Big Data and Machine Learning algorithms in cloud computing environments allows for real-time fraud detection with low latency. The methodology addresses the challenges of the CAP Theorem, eventual consistency, and massive data stream processing. The results demonstrate that decoupled architectures, when combined with AI predictive models, offer superior resilience and a critical response capability for mitigating financial risks on a global scale.