Looks pretty interesting, the main selling point for me would be the 'Fully reproducible' feature. I am wondering how you might achieve full reproducibility.
Yea I agree, especially combined with scheduling and dataset management. Setting up data science infrastructure can be such a pain.
While this is interesting, there is a couple of existing alternatives. Most notable would probably be JupyterHub, which brings Jupyter notebooks (or JupyterLab) to groups of users. The collaboration there is not realtime and there is no conflict resolution (last write wins), but you can share resources and files with your colleagues. Deploying it is also not super hard.
JupyterHub is a decent option, but it's still missing a lot of the features you need to really be a productive data scientist, such as IDE support (code intelligence, command palette) or data science related features (reproducibility, versioning or scheduling). It's also up to you to install it and maintain it — something you don't want to spend time on or you might not know how to do (not many data scientist are experts on managing infrastructure and keeping it secure).
Looks helpful. Interested to see how the product features stack up compared to Google Colab.
Google Colab is great for one-off experiments, but it's not something you can rely on for production workloads. You don't have any guarantees your machine will run for some time (it usually shuts down after 15 minutes), it's very difficult to import your existing code and version it properly.