Deploying the MLOps outer loop
The ML life cycle looks different for different use cases. However, the set of tools available in the Databricks platform makes it possible to automate as you like and supports your MLOps. The outer loop connects the inner loop products with the help of Workflows, Databricks Terraform Provider, REST API, DABs, and more. We covered automating the tracking process through MLflow Tracking and the UC Registry. The UC Registry is tightly integrated with the Model Serving feature and has a robust API that can easily be integrated into the automation process using webhooks. Each of these features can play a role in automating the ML life cycle.
Workflows
Databricks Workflows is a flexible orchestration tool for productionizing and automating ML projects. Workflows help the ML life cycle by providing a unified way to chain together all aspects of ML, from data preparation to model deployment. With Databricks Workflows, you can designate dependencies between...