Cloud-Native AI and Generative AI on AWS: A Systematic Review and a Proposed Unified Model

This fast development of cloud-native computing and Generative Artificial Intelligence (GenAI) has facilitated the creation of scalable and intelligent applications in various fields. But current literature and practice are full of individual studies on cloud-native machine learning pipelines and Generative AI systems, creating a fragmented architecture, haphazard governance, and more complicated operation in enterprise. Specifically, no combined architectural advice exists that combines principles of cloud-native computing with foundation model-driven GenAI processes into an integrated and consistent framework. The given paper provides a systematic literature review of cloud-native AI architectures, Generative AI models, and AWS-managed AI services to reveal major patterns of design, challenges, and gaps of the current practices. It was on this basis that we have developed a consistent cloud-native reference architecture to deploy and operate Generative AI workloads on Amazon Web Services (AWS). This work has the contribution of introducing a design-oriented AWSspecific reference architecture that applies standard cloud-native principles to unify traditional machine learning pipelines and Generative AI systems. Instead of concentrating on empirical performance measurement, the study suggests architectural guidance and practical design advice that are supposed to guide the organization in designing scalable, secure, and governance-conscious GenAI deployments on AWS. Future research will involve implementation-level verification and empirical evaluation of the proposed architecture in the real-life application contexts.