Rackspace MLOps Foundations

We help customers effectively operationalize machine learning (ML) models in production in a fraction of the time to accelerate innovation and ROI.

With Rackspace MLOps Foundations, we enable companies to accelerate operationalizing models in production to achieve the benefits of an automated cloud-based machine learning lifecycle management solution by reducing the typical 25+ step lifecycle down to 10 steps for a significantly faster delivery time.

MLOps Foundations, powered by Rackspace Technology Model Factory Framework, combines data preprocessing, model training, deployment and model monitoring across multiple environments with strong governance and security controls, to provide a foundation for successful model lifecycle management.

MLOps Foundations Features

Model Factory Framework is built using AWS services and open-source tools that enable rapid development, training, scoring and deployment of models. The MLOps Foundations solution can be tailored to specific workflows and business needs through customization around the Model Factory Framework.

Rackspace Model Factory Framework features

• CI/CD integration with automation and orchestration tools
• Tracking user interface
• Experiment tracking
• Project packaging for easy source control and reproducibility
• Register, package, train and deploy models from anywhere
• Support for major ML frameworks such as TensorFlow, SCikit-Learn, Spark ML, spaCy, PyTorch, etc.
• Real-time inference endpoints
• Scalable scoring
• Model artifact storage

ML model integration and pipeline development

• Refactoring of ML models
• Integration of reproducible ML pipelines
• Integration of reusable software environments for training and deploying models
• Automated triggering
• End-to-end QA test and performance checks
• Guidance to team collaboration and best practices

Getting Your Data Architecture Right

The Rackspace MLOps Foundation is delivered through a custom engagement. Rackspace Technology will work with you to determine your infrastructure needs and tailor your solution to meet your model development, operations and compliance requirements.

• Discovery Session: Start by working closely to understand your tooling, machine learning models and cloud landscape in order to define the problem space and develop a design solution.
• Design Decisions: Based upon the discovery session, a detailed design for the model factory is created using the tooling that your team is familiar with to decrease the maintenance efforts.
• Implementation of the Model Factory Framework: Once a design has been finalized, the team will begin implementing the solution. The Model Factory Framework provides a CI/CD pipeline for your models with a UI to make it easy to deploy and track the models.
• Integrate ML models: Once the solution has been deployed, Rackspace Technology will help you integrate your models and get them into production.

Delivery Project shall start no sooner than 30-60 days after Customer's acceptance of the Scoping Document, and the initial project under the first Scoping Document is anticipated to last for five weeks (to be confirmed in the Scoping Document)