Business Knowledge Series course

Presented by David Bronson is a Senior Linux Instructor and Red Hat OpenShift Delivery Certified Consultant with 25+ years of enterprise platform system administration experience. He currently serves as Senior Consultant with the Stone Door Group� and possesses extensive understanding of legacy platform technologies and migration to modern methodologies using RHEL administration, Docker/Kubernetes, DevOps, CI/CD, Jenkins, and containers. David co-developed this Stone Door Group and SAS Institute two-day ModelOps course, designed to teach SAS modelers how to transition to a DevOps mindset and implement those principles in ModelOps.

DevOps (development operations) principles have evolved over time within software development environments that create, update, and deploy high-quality software products. ModelOps (model operations) is a complementary approach that supports the additional unique challenges that are faced by those responsible for managing models within the model operations life cycle. This course presents conceptual and practical insights about how to get your production models up and running. Using hands-on practices, you learn how to perform ModelOps actions in the SAS Model Manager user interface as well as through the Jupyter Notebook.

Learn how to

• Define DevOps and related disciplines.
• Differentiate ModelOps from DevOps.
• Control the process of deploying models with the ModelOps approach.
• Manage models using SAS Model Manager and through “restful” APIs.
• Complete ModelOps tasks and milestones with Jupyter Notebooks.

Who should attend

Anyone who needs to know what ModelOps is; anyone responsible for managing the deployment of models in a production environment; operations team members that support model deployment and administration; and modelling team leaders wanting to get hands-on experience with ModelOps management software

Formats available	Duration
Live Web Classroom:	4 half-day session(s)

This course is relevant for anyone interested in learning about ModelOps. Attendees with experience in the Python programming language, Jupyter Notebooks, predictive modeling, and DevOps principles will find that background helpful, but it is not required.

This course addresses SAS Model Manager software.

Introduction to DevOps and ModelOps

• DevOps history, concepts, tools, and usage.
• How ModelOps practices build on lessons learned from the DevOps influence.
• ModelOps in the analytics life cycle.

Managing Projects and Models in SAS Model Manager

• Create projects.
• Upload model score code to the SAS Model Manager repository.
• Register Python models created in Jupyter Notebooks.

Program Interactions with Key Model Management REST APIs

• Authentication.
• Register models.
• Publish models.
• Create modeling projects.

Useful Capabilities for Preparing Models

• Clone the python-sasctl package using GIT.
• Package models from Python using python-sasctl.
• Create and manage models using Python SWAT (Scripting Wrapper for Analytics Transfer).