Unlike traditional dev workflows, ML workflows are difficult to run on a local machine (due to the lack of memory, more CPUs/GPUs, etc). This is why people often have to use remote machines (e.g. via SSH), or adopt one of the end-to-end MLOps platforms.

Using remote machines is not difficult but it is tedious and requires a lot of manual actions. Using MLOps platforms on the other hand automates the manual work but requires the use of an opinionated interface, which often kills developer productivity.

Imagine, if you could run your ML workflows the very same way as you do it locally, but they would actually run in the cloud. And you wouldn’t need to worry about provisioning infrastructure, setting up the environment, etc.

I’m excited to show you dstack, an open-source tool that does exactly that.

It’s a command-line utility that allows you to run any workflows while it provisions infrastructure, setup the environment, and copies code/data for you. No need to install or configure anything in your environment or cloud. Simply install the CLI and run it.

The launch blog post: https://mlopsfluff.dstack.ai/p/simplifying-the-mlops-stack

We’d love to hear your thoughts and ideas. I’ll be here to answer any questions you might have.