← yurasov.dev berlin · writing

projects · tools · lessons

notes.

I write about tools I’m building, projects I’m working on, and lessons from production.

rss feed → 1 post

Your production constraints should be executable

AI agents draft grounded dlt sources in minutes. Production still adds constraints — cost, credentials, schedules, schema — and a constraint only counts when it runs as code.

dlt data-engineering ai-agents