Hey everyone! I'm Emily, Sr AI Scientist at Future AGI. We work on AI reliability and I think a lot about orchestration because it's where so many things can quietly go wrong in a multi step AI pipeline. Prefect is something I follow closely. The question of where to put quality checks inside a workflow without slowing everything down is a design problem I find really interesting. The data pipeline world has thought about this more carefully than the AI world has. Happy to connect with people building AI pipelines here and thinking about how to make them more reliable and easier to debug.