Skip to content

Click in-app to access the full platform documentation for your version of DataRobot.

Deployment

To trust a model to power mission-critical operations, users need to have confidence in all aspects of model deployment. Many models never generate value because they never reach production. In the end, data scientists have to help deploy and maintain their models, which is costly and takes away from doing new data science.

With MLOps, the goal is to make model deployment easy. Operations teams, not data scientists, can deploy models written in a variety of modern programming languages like Python and R onto modern runtime environments in the cloud or on-premise.

There are two components to successful model deployment—creating the deployment and then monitoring and managing it. With DataRobot MLOps, you can create deployments in a variety of ways depending on the artifact you want to deploy.

Starting artifact Deployment method
DataRobot model Deploy a DataRobot model by selecting it from the Leaderboard and navigating to the Predict > Deploy tab. Alternatively, upload your model directly to the deployments inventory using the Add Deployment link.
Custom model Deploy a custom model from the Custom Model Workshop. Alternatively, upload your model directly to the deployments inventory using the Add Deployment link.
Remote model Upload the external model's training data directly to the deployments inventory to create a deployment. Alternatively, deploy an external model package. Once a deployment is created, you can instrument the MLOps agent to monitor the model and track prediction statistics. You can also upload historical prediction data after the deployment is created.
Historical prediction data Upload the training data for a model that made predictions in the past directly to the deployments inventory to create an external deployment. Add historical prediction data post-deployment.

Data overview

The tools and capabilities of every deployment are determined by the data available to them. These are the types of data your deployment can use:

Data Type Description
Training data Data used to train and build a model.
Prediction data Data that contains prediction requests and results from the model. Also referred to as inference data.
Outcome data Data that contains the actual values to compare to the prediction results (or Prediction Data) of the model. Also referred to as actuals.

Updated November 16, 2021
Back to top