Heyy, I am Bernardo Galvão. I am working on a - which is supposed to be a reimplementation of https://github.com/datarevenue-berlin/openmlops|OpenMLOps, but for on-prem servers because that is the requirement in my work., together with a deployed JupyteLab server. Anyway, there is a lot of heavy-lifting that ZenML has done, and of course, I find it better to leverage it and would love contribute to it :pleading_face:. I came here because I don't want to clutter the Issues board on GH. Here is a couple of points that probably should be on <#C01J2B14EFQ|feature-requests>:

• Do you have the Prefect Workflow/Dataflow Orchestrator in your roadmap? Their upcoming release, codenamed Prefect Orion, is bringing a powerful feature imo: conditional triggers. This would have the potential of emulating CI/CD and automatic retraining.
• As my requirement does not allow for a cloud-deployment, is there an alternative like packaging into a Docker Image and uploading to a registry?
I think it may be in my best interest contributing to ZenML in these topics. (Heck, even the Great Expectations Issue is looking good :eyes:)

@Bernardo Galvao great to have you here and what a cool project you have going. We'd love to contribute and collaborate.

• We dont have Prefect or Dataflow prioritized up but what we do have as high prio is to write docs on exactly how to create your own orchestrator! :smile: Maybe that might help? We are happy for a PR for either (Dataflow is way easier than prefect I feel)
• Not sure about this one, but if you see the cloud docs youll see the container registry is part of the stack and ZenML automatically packages your code and sends it to a remote orchestrator: https://docs.zenml.io/features/cloud-pipelines/guide-aws-gcp-azure
Great Expectations is a big demand! Also happy to have your help on that one!

1. yes that helps!
2. hmm, well, just packaging the code would be enough I think :slightlysmilingface:

Hi @Bernardo Galvao! Hows it going? Thought id jump in and ask you if youd had the chance to see ZenML develop

Unfortunately I am handling a lot of stuff. I should mention that my project is on-prem on Docker Swarm and is using Prefect, so it seems to me that our use case does not fit ZenML. :disappointed: