Clustering in Astrato via Snowpark, at Snowflake's Build ['22] event
I'm back with more Snowpark magic - clustering in Astrato, via snowpark. I'm able to re-configure and re-run models in real time, thanks to native Astrato's direct query and Snowpark for Python, from Snowflake.
Snowpark brings the promise of re-usable models in BI. No more throwaway python notebooks. Consumers are already able to configure models in real-time, using writeback. With a consumer-first approach, this frees up time from data scientists, reducing repeat requests and changes. This brings us dramatically closer to a low-cost self-service data science experience.
I built this for Build ['22], Snowflake's cloud dev summit event, in Tel Aviv last month. The event is described as packed with demos, AMAs, and hands-on labs created by builders, for builders. By far, it met expectations. Attendees were managers, developers, data engineers, and data scientists, mainly from Israel at one of the many thriving startups, or well-established large enterprises (some of which we already work with at Vizlib).