This thesis addresses the modernization of enterprise data integration by implementing a cloud-native Boomi iPaaS architecture to orchestrate data flows from Oracle environments, exploring how an advanced ETL framework can reconcile rigid corporate governance with the agility of an AI-driven ecosystem. The methodological approach centered on a hybrid deployment model leveraging the Boomi Atom, a lightweight and dynamic runtime engine acting as the execution plane to bridge on-premises data persistence with cloud-based management. This architecture utilizes bidirectional comparison logic and a database push-down strategy, offloading intensive computational tasks to the Oracle engine to optimize throughput and minimize JVM heap memory consumption.