How algorithms are deployed and monitored

- [Instructor] How are models deployed and monitored? … Model deployment is a stage of machine learning decisioning, … that is usually quite difficult to completely manage. … It is essential, but there are usually many issues … and questions, that appear earlier … in the development process, that delay this crucial stage. … There are also many risks that can be introduced … from the beginning stages, that can cause this stage … to drastically change. … At its core, deployment is taking the machine learning tools … from the development environment, and placing them … in a serving environment, where they're actually used … to assist decisions. … This photo of jumping out of a plane might be a bit … of a suitable metaphor, because at this point … of the project, where you start to take a number of chances, … and you start to see exactly what's happening … with the system. … In terms of its actual form, … there are many ways to deploy a model. … They can usually be encapsulated within a service, …