Show HN: A simple CLI for running ML jobs (or any jobs) with cloud compute (o-o.tools)

A new command line interface called o-o has been introduced to facilitate running machine learning jobs and other commands on ephemeral cloud instances. This tool allows users to easily build data and MLOps pipelines by stringing together multiple commands akin to local execution, but leveraging the power of cloud compute and storage. Key features include the ability to run any command in various environments defined by Docker images and machine types, track inputs and outputs for reproducibility, and manage data and source code securely in user-managed buckets. The significance of o-o for the AI/ML community lies in its flexibility and traceability, making cloud-based processing more accessible to developers and data scientists. The CLI currently supports Scaleway and Google Cloud, with plans for additional providers. Users can also integrate the tool with Git, allowing them to use source code from repositories within their cloud runs. While still in beta, o-o's ability to manage job workflows and maintain a history of runs—similar to version control systems—positions it as a promising resource for enhancing MLOps efficiency and reproducibility.