I’m working on a template for ML projects based on cookiecutter-data-science with a focus on reproducible research.
Right now, I’ve got a script to run jobs on a Docker container and the generated artifacts and metrics are logged on a Mlfow server. These jobs could be training a model or pre-processing some raw data for example.
What I want to accomplish is a way to run those jobs on the cloud: packaging a Docker image, upload it to a cluster and run the job.
Do you know any tool/platform I can use to do that?
I’ve checked out Cloud ML Engine, but I’m looking for a more generic docker-based way.
Thank you so much!