• Practically same approach than us at Mercadona Tech

  • Producer data marts and consumer data marts

  • Security: DBT and Terraform for roles!

  • Parse queries to monitor joins!!! That says something about your modeling

  • Metrics: Time to reliable insight

  • He doesn’t not see a need of different structures for ML and BI: DS would prefer curated data too (as long as your model is expressive enough).

  • Real time is overrated.

  • Analytics engineering not only is helpful downstream (analysts, DS) but also upstream to teams: if data models are so important in apps, having experts in data modelling will help your tram.

  • The default assumption (for teams) should be to expose everything.