How to productionize my MLPipeline to make live predictions with new data?

I’m satisfied with my MLPipeline performance. How can I productionize it such that it will generate a prediction when new data is loaded?