Overview
Syllabus
Intro
Pillars of Al Lifecycle - Datasets, Models...
and Pipelines
Define Pipeline with Python SDK
Argo Workflows
Kubeflow Pipelines with Tekton hits v1.0
Benefits of metadata and artifact tracking
Lineage Tracking
TensorFlow Extended-Using MLMD as metadata store
Kubeflow Pipelines v2 main goals
Machine Learning Metadata in v1
Pipeline Spec in v1
Intermediate Representation in v2
New Orchestration Controllers
Components
Abstraction Layer for Orchestration Engines
Abstraction Layer Benefits
Abstraction Layer Features: Execution Client
Abstraction Layer Features: Execution Spec
Summary
References
Smart Runtime
Taught by
Linux Foundation