Master Efficient ML Model Deployment: Optimize, Scale, and Control Your BERT Transformer API with AWS EKS, Ray Serve, and K8s HPA