Three teams were transforming the same payment data in four different ways. We consolidated it onto one Databricks lakehouse in 14 weeks, without a freeze on reporting.
A digital health provider's nightly reporting load kept running into the working day. Most of the time went into reloading data that had not changed.
Support engineers were spending around 25 minutes per escalated ticket searching five systems. Our first prototype answered well, but also exposed data across customers. We fixed the permissions first.
* Some entries are illustrative examples and will be replaced with published client work.