The integration of cloud-native architectures into artificial intelligence (AI) workflows has revolutionized the deployment, scalability, and efficiency of machine learning (ML) solutions. This study explores the design and evaluation of cloud-native ML models within cloud-first infrastructures, emphasizing their performance, cost-effectiveness, and scalability. Leveraging platforms such as AWS, Google Cloud, and Microsoft Azure, the research investigates data pipeline efficiency, model training metrics, inference performance, and economic viability. Statistical analyses reveal consistent accuracy, precision, and recall across models, with distinct trade-offs in resource utilization and latency between batch and real-time inference methods. The findings highlight the transformative potential of cloud-native ML in optimizing AI-driven decision-making, while identifying challenges such as resource allocation and cost management. This study serves as a foundation for advancing AI applications in cloud environments, offering insights for organizations to achieve greater agility and efficiency in AI deployment.
Gupta et al. (2024) studied this question.