Skip to content

Comment on A Python-first data lakehouse

Comments

Rolling a notebook out to a service rapidly is an attractive idea -- but, as mentioned, has security implications -- I can add that there are also a host of monitoring implications as well -- service quality & continuity, model quality etc.

You mean on the data side? Data access in the example (and in real-world) is mediated by production-grade Iceberg compatible catalog, sandboxed changes, and full auditability trail (https://docs.bauplanlabs.com/en/latest/concepts/git_for_data...). Or do you mean something else?

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.