The rapid adoption of microservices architecture has increased the need for efficient and reliable software design techniques. UML class diagrams are commonly used to represent system structure during the early stages of development. However, existing UML-to-code generation tools directly convert designs into code without validating design quality, often resulting in tightly coupled and poorly structured microservices. This paper proposes an AI assisted UML-driven validation and optimization framework for automated microservices code generation. The proposed system analyzes UML class diagrams using a combination of rule based validation and machine learning techniques. It identifies design issues such as tight coupling, missing relationships, and improper naming conventions. Based on the detected issues, the system provides intelligent suggestions for optimizing microservice boundaries and improving overall design quality. The optimized UML model is then used to generate microservices-based full-stack application code. The proposed approach improves scalability, maintainability, and reduces manual effort in software development.