Notebooks have been an unexpected innovation in how software is written. They have lowered the bar for writing scripts, unlocking huge amounts of scientific productivity.
But are they going to make it into 'production software'?
I'm going to introduce Treebeard's notebook CI framework and how we built it with Papermill, Repo2Docker, and GitHub Actions.
In this talk I will describe the evolution of notebook use-cases and the great potential of investing in adjacent tooling.
Given my perspective from working in devops on large websites, I'd like to impress that the Jupyter project has discovered not just a great product for exploratory data science, but for software engineering more generally.
Finally I'd like to introduce the problem area we are working on -- continuous integration, in order to accelerate what we see as a beneficial progression in software.
I will introduce you to some open source projects: