Container

The container is a figurative structure used to keep track of where your model(s) is being used. Within RTL we use to deploy many ML models since the data science team wants to improve the user-experience. To keep track of where the models are used, we created the container entity.

Content

Each container has three important parameters which two are mandatory to have in order to create it. The properties are:

1. PublicationPoint. The system where the model it is deployed
2. Campaign. In which part of the system the model is used
3. Models. The list of models used for this particular combination of publicationPoint and campaign

Let's give an example why of this choice. Let's assume that you have a social-network website called Proton. You website is composed by many sub-parts. Some of this parts are profile, photos, news-feed, etc. To encapsulate the usage of the models, the container object would have the following values:

• PublicationPoint = "Proton"
• Campaign = "Photos"
• Models = ["face_recognition","object_tagging"]

We can then create another container. For example:

• PublicationPoint = "Proton"
• Campaign = "news-feed"
• Models = ["collaborative-filtering"]

And so on. In this way we are able to keep track of the usage in a simple way. Note that the Models values are the name of the Models that are created in a separate manner. With this clear separation between container and model we can re-utilize the models in many containers.