Youtor
DataRobot 90 Compliance Documentation MLflow
>> YOUR LINK HERE: ___ http://youtube.com/watch?v=LvSgQrgyv64
https://www.datarobot.com/platform This video from the March 2023 Launch Event for DataRobot 9.0 features Senior Director of Product Management Brian Bell giving a tour of two new capabilities in ML Production related to model documentation. • Learn more at: • https://www.datarobot.com/platform/ne... • https://docs.datarobot.com/en/docs/ml... • Speaker • Brian Bell, Senior Director of Product Management, DataRobot • Contents • I'm Brian Bell MLOps Product Lead at DataRobot. Today I'm going to dive into the world of model governance and how that affects all enterprise AI today. • Compliance docs save data scientists lots of time by automating the work of model documentation with the click of a button. Now let me show you how easy it is to document your models, including external ones, as we've added in 9.0. Now we're going to bring over all of our critical stability checks and experiment run history that was stored in MLflow when we built our model. Here you can see my run info in MLflow with a description, training parameters, metric scores, tags, and artifacts. For example, I really want to add this chart to my documentation. We want to bring these key values into DataRobot to use inside of a Compliance Document and generally centralize all of our model metadata across all of our building scenarios. Here I have a model in my registry that I want to populate with imported artifacts and metadata from MLflow. • So now we'll use our MLflow integration from the command line on the right to import all this information and attach it to this model inside of the DataRobot AI Platform. Here you can see it automatically syncing all this model information and adding it to DataRobot in seconds. Here's our chart artifact that we saw in MLflow now attached to our model in DataRobot. Now we can run a new template that adds these MLflow key values into a compliance report. Just one click and DataRobot pulls everything together. As you can see the compliance doc has those custom values that we imported from MLflow. So now all of your custom model scenarios descriptions charts and other key content can be added in a repeatable way to enrich your Docs. We're super focused on helping data science and model risk collaborate effectively to document models fast. Our repeatable workflow helps AI Builders save time and increase quality. • Explore the full DataRobot AI Platform at https://www.datarobot.com/platform • Stay connected with DataRobot! • ► Blog: https://blog.datarobot.com/ • ► Community: https://community.datarobot.com/ • ► Twitter: / datarobot • ► LinkedIn: / datarobot
#############################
