The fast growing need for AI applications leads to the employment of complex Artificial Neural Networks (ANNs) in cloud platforms. An approach consisting of several layers is explained which employs ML as well as DL algorithms combined with autoscalers and load balancers to deliver the most important traits of efficiency and scalability for models on cloud. The developed system uses a dynamic autoscaling approach that re-distributes resources according to the current performance of the system. The system developed makes use of dynamic autoscaling technique that adjust resources according to the current load of the system. Load balancing techniques are then put into place to ensure that processing is done uniformly across resources, hence decreasing latencies and maximizing throughputs. Study scrutinizes problems experienced during deployment of ANNs including tasks with dynamic workloads, limited resources, uncertain user demands. The research also discusses the design and implementation of the ANN framework, providing greater detail into the algorithmic decisions made and their effect on system efficiency.