Deploy code note models

This is a new MLOps design paradigm in which instead of promoting trained ML models across accounts, the whole underlying code is promoted and the model is re-trained with data in the new account

Workflow including tooling account:

• Commiting to the build repo (in tooling acc), triggers the build pipeline (in tooling account)
• Build pipeline triggers SM pipeline (in dev) which ends with a model registered to Model Registry Dev (in tooling)
• Approving the model triggeres the dev deploy pipeline (in tooling) which creates the infrastructure in the dev account
• Manually approving the build pipeline step (in tooling), triggers the SM pipeline for staging (in stg account)
• The SM pipeline registers a model in Model Registry Staging (in tooling account)
• Approving the model version triggers the staging deploy pipeline (in tooling) which craetes the infrastructure in staging account (Repeat for Prod)

Resources in tooling:

• 2 Code repos
• 3 Deploy Pipelines
• 3 Model Registries
• 3 Event triggers (trigger deploy on model approval)
• 1 Build pipeline