Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Duration 14 hours
Course Outline
Introduction to Kubeflow
- Examining the goals and architecture of Kubeflow.
- Overview of key components and the broader ecosystem.
- Exploring deployment strategies and platform features.
Interacting with the Kubeflow Dashboard
- Navigating the user interface.
- Administering notebooks and workspaces.
- Connecting storage solutions and data sources.
Foundations of Kubeflow Pipelines
- Structuring pipelines and designing components.
- Writing pipelines using the Python SDK.
- Running, scheduling, and monitoring pipeline executions.
Training ML Models on Kubeflow
- Strategies for distributed training.
- Utilizing TFJob, PyTorchJob, and other operators.
- Managing resources and autoscaling within Kubernetes.
Serving Models with Kubeflow
- An introduction to KFServing / KServe.
- Deploying models using custom runtimes.
- Overseeing revisions, scaling, and traffic routing.
Orchestrating ML Workflows on Kubernetes
- Versioning data, models, and artifacts.
- Incorporating CI/CD into ML pipelines.
- Implementing security and role-based access control.
Production ML Best Practices
- Crafting robust workflow patterns.
- Establishing observability and monitoring systems.
- Resolving common Kubeflow challenges.
Advanced Concepts (Optional)
- Configuring multi-tenant Kubeflow environments.
- Scenarios for hybrid and multi-cluster deployments.
- Expanding Kubeflow with custom components.
Conclusion and Future Steps
Requirements
- A solid grasp of containerized applications.
- Hands-on experience with basic command-line operations.
- General familiarity with Kubernetes principles.
Target Audience
- ML practitioners.
- Data scientists.
- DevOps teams new to the Kubeflow environment.
Testimonials (1)
I like the Kubernetes part of the training as I had some tasks with K8s pods and ports and it was nice to get a better understanding now. Thank you!