ZenML workshop demonstrating agentic pipelines with dynamic execution and deployment.
Research pipelines using ZenML dynamic pipelines and pydantic-ai agents.
Basic research pipeline: plan research topics, execute searches in parallel via .map(), synthesize final report.
Extends 01 with prompt tracking as artifacts for lineage and versioning.
Extends 02 with query validation step using conditional branching.
Standalone examples demonstrating ZenML dynamic pipeline features:
| File | Concept |
|---|---|
parallel_steps.py |
Parallel step execution with .submit() |
parametrize.py |
Runtime step parametrization via YAML config |
product.py |
Cartesian product mapping with .product() |
unmapped.py |
Passing unchanged inputs to mapped steps |
unpack_tuples.py |
Unpacking multi-output results from mapped steps |
cd dynamic_pipelines
uv sync --dev
uv run zenml init
uv run python 01_pipeline.py
...Weather agent pipeline deployable to ZenML Pro.
cd deployment_example
uv sync --dev
uv run zenml init
uv run zenml login
uv run python weather_agent.pyuv run zenml pipeline deploy weather_agent.simple_weather_pipelineApache License 2.0