A comparison of Great Expectations, Soda, dbt tests, and anomaly detection, and a layered design that uses each where it fits.
The case for generalists owning end-to-end data workflows with agents, the counterargument, and how to make the transition work.
How dbt incremental materializations map to Iceberg operations, and the configuration, predicates, and maintenance that keep them healthy.
Disaster recovery for Iceberg across four tiers: snapshots, object versioning, catalog backup, and cross-region replication.
