Development plan and contribution

[yxue-kabam]

Description
I have built our ML infrastructure with airflow and kubernetes. As a centralized scheduler, airflow has a few scenarios where airflow does not work well. For the next iteration of our ML workflow and scheduler, I'd like to invest in cloud-native workflow engine (basically argo). I'm actively exploring with this library and am wondering about the longer term development plan?

Do you take external contribution at the moment? (I suspect not at the moment, but would like to know when you would call it a V1 and go from there?)

Additional context

[brtasavpatel]

@yxue-kabam I am also looking to do the same, I tried making airflow work with k8s executor but it was real pain to get it working and managing it so now have moved to argo. I would be happy to contribute as well.

Also, sorry for shameless plugin to my question #12 here. do you have any insights into this one?

[CermakM]

Thank you for the interest, guys! Appreciated.

I've been working closely with the Argoproj and we're in the process of migration of https://github.com/CermakM/argo-client-python under Argoproj Labs to give it a more official status and attract more contributors. I suspect this repository will become part of that organization, too so that we could boost the development, but it is yet to be discussed.

I intend to release v0.1 soon once I get some of the current issues fixed, but I have been quite short on time recently, so my plans have gotten a bit delayed, unfortunately. Contributions are and always will be welcome! :) so are the ideas for further improvements and proposals for enhancements.

[hadim]

@yxue-kabam do you have feedbacks about using Argo and this library to build an ML cluster?