MLOps with Red Hat OpenShift: A cloud-native approach to machine learning operations

MLOps with Red Hat OpenShift: From Platform Setup to Production Models 💻

Machine learning initiatives often run into trouble not at the model-building stage but in the space around it: environments, data versions, deployment paths, monitoring, and repeatable operations. MLOps with Red Hat OpenShift addresses that operational layer through a cloud-native approach. Ross Brigoli and Faisal Masood, both experienced software and cloud architects, guide readers from the foundations of MLOps and OpenShift into the practical work of provisioning, configuring, and running machine learning workloads on Red Hat OpenShift.

The book is structured in three parts. It begins with the concepts behind MLOps, OpenShift, operators, ROSA, and OpenShift Data Science. It then moves into provisioning an MLOps platform in the cloud, covering AWS preparation, ROSA installation, partner software, and Pachyderm. The later chapters turn to operating ML workloads: managing training workflows, deploying models as services, monitoring, logging, optimizing cost, and building a face-detector system on the Red Hat ML platform.

Inside the MLOps Lifecycle ⚙️

Rather than treating tools as isolated topics, the book connects them across a working lifecycle:

• Introduction to MLOps and OpenShift, including operators, ROSA, and OpenShift Data Science
• Provisioning an MLOps platform in the cloud, with AWS and ROSA setup
• Building models with Jupyter Notebooks, ML frameworks, GPU acceleration, and custom notebook images
• Managing data versioning and training pipelines with Pachyderm and OpenShift Pipelines
• Deploying models as services with Seldon, autoscaling, rollbacks, canary releases, and endpoint security
• Operating workloads with Prometheus, Grafana, inference logging, and cost optimization
• Bringing the pieces together in a face-detector project using the Red Hat ML platform

Tools in Context, Not Just in Theory 🧠

OpenShift Data Science, Pachyderm, Seldon, Prometheus, and Grafana appear as parts of an MLOps architecture. Readers see how platform decisions affect model training, deployment, and day-to-day operations. The material is especially relevant when a team needs a repeatable path from notebook experiments to services that can be monitored, scaled, and updated safely.

Who Will Benefit

Data scientists, ML engineers, platform engineers, DevOps and SRE practitioners, and architects will find the book aimed at their work. It assumes an interest in cloud-native infrastructure and machine learning operations rather than offering a first course in Python or statistics. The emphasis stays on the operational practices that help ML systems run reliably on OpenShift.

A Practical Path to Operational ML

If your ML projects need a stronger platform foundation, MLOps with Red Hat OpenShift offers a structured route through the OpenShift tooling and workflows that support the machine learning lifecycle. It is a focused technical guide for readers who want to understand how cloud-native infrastructure and MLOps practices meet in real deployments.