MLOps Foundation is a two-day, hands-on course that teaches attendees how to build enterprise-grade machine learning platforms and infrastructure. Day one opens with an overview of the discipline, then moves into machine learning service management. This covers the operational concerns of keeping models running once they leave a notebook. Attendees examine ML architecture and tooling choices before applying a CI/CD approach to machine learning workflows. This treats model training and validation like any other automated pipeline that gates a release on repeatable, measurable results.
Day two turns to deploying and running that infrastructure. The class starts with a Kubernetes overview, then layers on orchestration with Kubeflow to coordinate multi-step training and deployment pipelines on the cluster. Attendees work hands-on with Ray for scalable, distributed computation across a pool of workers. They finish by serving models in production with Seldon Core. This involves wiring up an inference endpoint that can be monitored and scaled like any other service.
Participants leave with practical exposure to the architecture principles, deployment strategies, and prevailing technologies that power modern machine learning processes. This spans both on-premises and cloud environments. They also gain hands-on lab experience across each tool covered in the two days.
Who Should Attend
AI/MLOps, DevOps and DataOps engineers; Data Scientists/Engineers, Developers, IT and QA Staff, Technical Managers
What Attendees Will Learn
Upon completing MLOps Foundation, participants will be able to:
- Understand the fundamental concepts of incorporating machine learning into production processes
- Gain hands-on experience with frameworks for training, deploying, serving and monitoring machine learning models
- Learn how to adopt a CI/CD approach to machine learning workflows
- Become familiar with the architecture of machine learning processes and services
- Gain exposure to prevailing technologies in the rapidly developing space of MLOps
- Learn how to implement scalable solutions to support machine learning workflows
- Deploy machine learning models using various strategies, including on-premises and cloud deployment
Prerequisites
Participants should have some programming skills (preferably in Python), familiarity with Linux and basic command-line skills.