Keeping your code organized and versioned is of obvious importance to software developers. It saves us precious time. The machine learning (ML) field is only beginning to catch up. This stems from fundamental differences in workflows, requiring different tools.
In this talk, we will review these counter-intuitive differences and the state of the ML world today. We’ll introduce DVC, an open source command line tool that complements Git and enables organizing and versioning of machine learning projects, in order to fill the gaps required by the ML workflow. We will show a case study, and attempt to encourage the growth of an open source ML community in Israel.