More and more data are being generated or made available in cloud. More and more enterprises have investment in cloud. There is the need for data scientists and developers to build and train AI models in cloud to utilize the benefits of cloud, for example, elasticity, scalability, and pay-as-you-go economy. Meanwhile, mainframe is still the main transaction processing engine for lots of enterprises for unbeatable performance, reliability, and security. Some use cases require large-scale real-time low-latency model inference. IBM AI Telum processor on z16 can help clients achieve that and allow clients to make on-the-spot decisions while the transaction is happening. Combining the power of cloud and mainframe can bring client the benefits of both worlds. There is the need for developers and data scientists to be able to develop and test the AI models in cloud and deploy the models to mainframe for real-time inference.This document demonstrates how to build a sample model using the services in Cloud Pak for Data in IBM Cloud and how to deploy the saved model to mainframe for real-time inference.