Storage

A risk warehouse differs from a general one in what it has to prove: the same exposure has to reconcile across IFRS 9, internal ratings and regulatory reporting, and every figure has to trace back to the record that produced it. At Gini we build that warehouse on the platform that suits you, whether that is Databricks, Snowflake, BigQuery, Redshift or SQL Server, with audit and lineage designed in from the start rather than added once the first reconciliation fails.

Yes. The choice follows your estate and what your team already runs. Where a migration is genuinely worth doing, usually because analytic tooling you need is out of reach where the data currently sits, Gini says so and costs it. Where it is not, we build on what you have.

At Gini we treat them as a design decision rather than a running cost. Partitioning, clustering and the boundary between stored and computed results are set at build time, and query patterns are reviewed against the bill after the first month. Most overruns we are asked to fix come from a small number of unpartitioned tables scanned repeatedly by scheduled jobs.

The schemas, the transformation code, the tests and the deployment configuration, version controlled in your repository and running under your credentials on your own platform subscription. Gini walks your engineers through it before handover, because the point is to leave you able to extend the estate without calling us.