Get an overview of MLOps (Machine Learning Operations) and discover how Kubernetes and Kubeflow can streamline your ML workflows. This talk will introduce the core principles of MLOps, highlighting its role in ensuring reliable, scalable, and efficient ML deployments.

We'll then dive into Kubeflow, a powerful Kubernetes-native platform purpose-built for MLOps. You'll learn about Kubeflow's key components, including:

Pipeline Orchestration: Building and automating end-to-end ML workflows
Notebooks: Interactive development environments for experimentation
Model Serving: Deploying models for real-time or batch predictions
Training: Scaling model training on Kubernetes
Hyperparameter Tuning: Automate hyperparameter optimisation to find the best model configuration and improve model performance with Katib
Security & Access Control: Integration with Kubernetes RBAC

Key Takeaways:

Understand the core concepts of MLOps and why it's important
Learn how Kubeflow simplifies common ML challenges
Gain insights into Kubeflow's key components and how they work together
Get tips for starting your own MLOps journey with Kubernetes and Kubeflow.