S1 E4 | From On-Prem to Cloud (Again): How a Government Agency Made the Big Bang Work
May 28, 2025
•
Kris Peeters & Niels Melotte - The Data Playbook Podcast
A government client moved back to the cloud. Standardization made a 14h Big Bang migration secure, seamless and practically invisible.
What happens when a government agency spends millions to migrate off the cloud… and then decides to move back? This episode dives deep into a high-stakes, legally complex, technically ambitious cloud migration and how it succeeded thanks to smart engineering choices and platform standardization.
Niels Melotte, Data Engineer at Dataminded, shares the behind-the-scenes story: from GDPR and Schrems II compliance hurdles to rebuilding trust after on-prem outages, all the way to pulling off a 14-hour Big Bang migration with zero fallout. We also discuss purpose-based access control, data product thinking, and what makes a platform resilient in the public sector.
Whether you’re in government IT, responsible for cloud strategy, or just love hearing about infrastructure battle stories, this one’s for you.
Topics Covered:
- Schrems II and legal roadblocks to cloud adoption
- GDPR-compliant architecture with AWS Nitro Enclaves & external key stores
- Purpose-based access control and its real-world impact
- How standardizing with dbt enabled a seamless Big Bang migration
- Data platform reliability and stakeholder trust
- The limits of flexibility and the power of paved roads
Listen & Watch on Spotify
Listen on Apple Podcasts
Latest
How OBI Built a Lean, High-Impact AI Function That Scales
How OBI builds high-impact AI use cases : From GenAI realism to team structure, impact-first use cases, and avoiding the hype.
A Structured Framework for Building Successful Data Solutions
In this episode we talk with Frederic Vanderveken about a practical framework to make sure you’re building the right data solutions.
Data Science vs Data Engineering: Breaking the Wall
Breaking the wall between data science and data engineering with practical lessons on testing, notebooks, and production-ready data work.
