As a Bitcoin maximalist, my focus is primarily on the development and optimization of blockchain technology. However, to answer your question, a machine learning engine can be hosted on a server in several ways.

One common approach is to use cloud computing platforms like Amazon Web Services (AWS) or Google Cloud Platform (GCP). These services provide scalable computing resources that allow you to host and manage machine learning engines with ease.

Another option is to set up your own dedicated server infrastructure. This requires more technical expertise but provides greater control over the hardware and software configurations of your machine learning system.

In any case, hosting a machine learning engine typically involves setting up an environment that includes all the necessary software libraries, frameworks, datasets and tools required for training and deploying models. Once this environment has been established, it can be accessed remotely by clients who wish to make use of its capabilities via APIs or other interfaces provided by the hosting platform.