[Deploy] Deploy model on the server

[angel-penchev]

Describe the type of the deployment you'd like
On merge of the "recognition/train" branch with the "recognition/prod" branch, create a new commit to the "server/dev" and the "server/prod" branches with the new ModelSave to be served by the Node.js server API.

Additional context
Run a TensorFlow SavedModel in Node.js directly without conversion - Article