Hello folks, Does anyone know how hera (the official way to use Argo Workflows - https://github.com/argoproj-labs/hera) compares w/ Metaflow using Argo Workflows as the backend?

I think the difference is pretty fundamental: • Metaflow is a stand-alone framework for building ML/AI systems with many batteries included: it manages state and the data flow (through artifacts), all dependency management, versatile compute, observability, name spacing, and local development etc out of the box • Hera is a Python DSL for Argo Workflows. It doesn’t do anything that helps you write systems in Python (no deps management, no state management etc) - it just executes Python scripts that you put in Docker images manually If you need a thin Python DSL for argo, Hera is definitely the right tool for the job. If you need anything more than that, Metaflow has a much better chance of delivering what you need

Perfect! Thanks @victorious-lawyer-58417.

A >> [B, C] >> D  # Define execution order

This is an interesting syntax