How does dstack compliment Jupyter notebook and other tools?

dstack is a platform for building internal in-house data applications, e.g. for quick prototypes of ML models, or for advanced analysis or prediction in business units as an addition to normal business intelligence tools (Tableau, Power BI, QuilkSense, Metabase, Superset, etc).

The purpose of dstack's applications is not to replace Jupyter notebooks (dstack can be used from Jupyter notebooks) but rather to build in-house data applications for non-technical colleagues within the company.

Which application outputs are supported?

It supports pandas, matplotlib, Plotly, Bokeh, and Seaborn.

How does dstack help with building ML applications?

The dstack framework offers an ML Registry where you can push your models and later pull them to be used within applications. It supports Scikit-learn, Tensorflow, and PyTorch models.