ML Ops

MLOps Overview

The MLOps section provides a centralized interface where all your AutoML Objects used in Reasoning Flows are listed and managed.

From this dashboard, you can:

• View all AutoML Objects across projects.
• Identify their current status (Active / Inactive).
• See which engine they are running on (e.g., dk-automl-engine-gpu).
• Check who created or last updated each object.
• Use quick actions to open or manage AutoML Objects directly.

Each listed item gives you a snapshot of key details such as:

• Project name
• Associated dataset
• Creation date
• Owner
• Operational status (e.g., Active)
• Shortcut menu for object-specific actions.

---

Object Actions

Each AutoML Object includes a quick-action menu (⋮) with the following options:

• Open Object – Opens the AutoML Object details view, allowing you to review its configuration and metadata.
• Manage Models – Opens the models view, where you can monitor, deploy, or remove models associated with this object.

---

AutoML Object Details

When you open an object, you’ll see the AutoML Object Information panel containing:

---

AutoML Object Models

Below the object information, you’ll find a list of all models associated with the AutoML Object.

Each model entry displays:

• Model name (e.g., mdl_forecast_2024)
• Deployment status (e.g., ✅ Deployed or 🟣 Trained)
• A context menu (⋮) with model-level management options.

---

Model Actions

From the model options menu, you can:

• Deploy Model – Makes the selected model active and available for use within its Reasoning Flow.
• Remove Model – Deletes the model from the AutoML Object, freeing resources and keeping your flow organized.

Additionally, you can:

• Train a new model using the Train Model button.
• Close the panel to return to the MLOps main dashboard.

---

Summary

The MLOps provides a unified workspace to:

• Track, configure, and manage all AutoML Objects.
• Monitor model lifecycles (training, deployment, removal).
• Streamline Reasoning Flow operations for efficient model governance and maintenance.

All tools are conveniently located in one place for faster, more transparent management of your AI automation flows.