Pick the migration you're actually facing.
Concrete source-to-target shapes the team carries a codified rule library for. Each page covers the dialect translation, the architectural decisions, and the gotchas that typically eat the timeline — so you can scope the engagement before the first call.
Snowflake → BigQuery
Credit economics to slot economics. DDL translation, CLUSTER BY mapping, JS UDF rewrites, Time Travel replacement, Iceberg-vs-native storage call. Fixed-fee, reconciled object-by-object.
Redshift → Snowflake
The most-asked pair. DISTKEY / SORTKEY → CLUSTER BY, IDENTITY column handling, SUPER → VARIANT, WLM queue economics → warehouse sizing, Spectrum tables → external stages.
BigQuery → Databricks
GoogleSQL ↔ Databricks SQL parity audit, GBQ ML → MLflow rewrite, slot reservations → SQL warehouses + clusters, Unity Catalog adoption, Iceberg or Delta target call.
On-prem → Cloud
Teradata, Hadoop, on-prem Oracle, SQL Server, Informatica / SSIS — anything to AWS, GCP, Azure, or a managed warehouse. Inventory, dependency mapping, wave plan, parallel run, cutover.
Multi-cloud rationalization
Three cloud bills, two storage buckets, and a Kubernetes cluster you cannot quite explain. We map the actual workload-to-cloud topology, name what should stay split and what should collapse, then ship the consolidation in fixed-fee waves.
Cloud cost reduction
When the bill outpaced the platform team. Rightsizing, commitment correction, idle cleanup, egress, Kubernetes spot, warehouse query rewrites — sequenced as a 90-day plan with savings per lever.