ML models deployment on edges

Dear Community,

from a central cloud Cumulocity instance with MLW do we have the possibility to deploy ML models towards Cumulocity Edges remote installations?

As I am able to schedule monthly retraining of the model I need also to automate remote deployment of the model which need to run at the edge.


Hi Mauro,

I think currently what you want to achieve would not work with the standard implementation of the connected-edge and the used agent on edge side.

However I see how this could be realized with if you have a dedicated thin-edge installation on every edge that is connected to that central instance. You could use an own software management plugin on edge side.