Concept | The Govern node’s role in the Dataiku platform#
Within the Dataiku platform, the Govern node is the centralized place dedicated to managing the governance of analytics, models, and agents at the enterprise level.
The Dataiku platform suite#
If you have a background in MLOps with Dataiku, you’ll already be familiar with Dataiku’s family of integrated nodes. If not, note that Dataiku, as an end-to-end platform, includes the following environments for different tasks of the AI lifecycle:
Node |
Purpose |
|---|---|
Design |
A sandbox environment for building analytics, models, and agents |
Deployer |
A place for managing the deployment of project bundles and API services |
Automation |
A production environment for batch deployment of project bundles |
API |
A production environment for real-time deployment API services |
In the simplest terms, you could summarize a typical production workflow in Dataiku as:
Build a package (such as a project or model) in the Design node.
Publish it to the Deployer.
Deploy either A) a bundle to an Automation node or B) an API service to an API node.
See also
If you’d like to gain a stronger understanding of these concepts, explore the MLOps Practitioner learning path.
The Govern node#
These components work together as a complete ecosystem for building, deploying, and monitoring analytics, models, and agents. At the same time, to successfully scale their AI journeys, many organizations recognize the need for a dedicated layer of governance as part of their technology stack.
The Govern node is an additional component of the Dataiku platform that directly addresses many of the challenges of AI governance.
To get started, many organizations first adopt the Govern node to secure their MLOps workflows. They do this by positioning the Govern node as a gateway to deployment infrastructure. Instead of allowing packages to proceed directly from development to production environments, these organizations route deployment decisions through the Govern node.
Without an approval from a sign-off process in the Govern node, packages can’t move from a development environment to a production environment. To use Dataiku terminology, a project or model can’t move from the Deployer to an Automation or API node.
Tip
While securing MLOps may be your organization’s starting point, the Govern node’s role can extend across the AI lifecycle. For example, this could include pre-qualification in ideation and progress documentation during development.
You’ll get a chance to walk through this process in Tutorial | Governance lifecycle.
Next steps#
One out-of-the-box capability of the Govern node is its ability to sync metadata of items from connected Dataiku nodes. Learn more in Concept | Surfacing Dataiku metadata in the Govern node.