databricks-model-serving

Deploy MLflow models to Databricks Model Serving endpoints by running Python scripts that use the Databricks SDK to provision scalable REST APIs, enabling secure production access to your machine learning predictions.

Yes, you can deploy LangChain AI agents to Databricks endpoints. The skill handles deployment for GenAI agents built with langgraph and langchain, creating queryable REST API endpoints for your applications.

Yes, Unity Catalog is required to use Databricks Model Serving. You must have Unity Catalog and Databricks Model Serving features enabled in your environment to manage deployments and endpoint security.

The best way to query a deployed ML model on Databricks is by sending requests through the Databricks SDK to your active Model Serving endpoint, retrieving predictions or responses directly in your Python scripts.

Manage and update deployed endpoints in Databricks using the Databricks SDK to monitor status, modify configurations, and apply updates to your MLflow models and AI agents serving traffic.

Yes, this skill supports deploying custom Python models. You can deploy MLflow models, custom Python models, and GenAI agents to scalable endpoints using the provided deployment scripts.