In this marimo-hosted livestream, I gave the first technical preview of Polars Cloud and showed how Polars can scale from a single node to a distributed cluster.
Key Points
Polars and its growth
Polars has come a long way since its early days. Download numbers are high and adoption keeps accelerating. More and more teams are asking whether they should switch from Pandas.
Polars Cloud features
The main focus of the session is Polars Cloud: a platform that lets you run distributed queries across multiple machines without rewriting your code. We are working towards a generous free tier so anyone can experiment with it.
Streaming engine and lazy execution
I explain how the streaming engine works: it processes data in chunks rather than loading everything into memory at once. Combined with lazy execution, Polars optimizes a full query plan before running anything. That combination is what makes the performance at scale possible.
Live demo
I run a query live and compare the results against Pandas. The lazy API does the heavy lifting, and the performance difference is significant.
Future and community
I close out with a look at what is coming next and how you can get involved through Discord. Polars Cloud is still early, and we are actively looking for feedback.