b
Re: live model serving
As others have mentioned in this GH issue; it is possible to do model deployment as an "end" step in a workflow.
Our team is curious if abstracting deployment to common live model serving targets such as Sagemaker is on Metaflow's near term roadmap at all? We noticed MLFlow does this
nvm found this!
https://outerbounds.com/docs/deploy-with-sagemaker/