Approach
How we build data platforms that stay yours.
Three principles shape every design. Underneath sits a lakehouse: the technology that separates your data from the engine on top.
Approach · Principles
Three principles we stand for
Always open formats
Your data lives in open, standardised formats on ordinary storage. Any compute engine can read it: today's vendor's and tomorrow's.
Exit strategy built in
Every platform we deliver includes a tested exit drill: the same project and tables, running on a second platform. The runbook for leaving is ready from day one.
Pragmatic and tailored
Every dataset gets the level of control that fits. We classify first, then place each workload on the level of sovereignty it actually requires.
Approach · The technology
What is a lakehouse?
A lakehouse combines the cheap, flexible storage of a data lake with the structure and reliability of a data warehouse. Your data sits in open file formats on ordinary object storage. Different engines read and write the same tables, on one shared copy, and the keys stay in your hands.
Data warehouse
Structured and fast for reporting. Storage and compute live in one product from one vendor, in its own format. Switching calls for a full export and rebuild.
Data lake
Cheap and flexible: all data as files in object storage. For reliable analytics, it needs table structure and transactions on top.
Lakehouse
Files in object storage, with an open table format (Apache Iceberg) and a catalog on top. The data behaves like a warehouse, while storage and compute stay separate. Your data stays put when you switch engine or vendor.
That is why a lakehouse is the foundation of a real exit strategy: the data stays put, only the engine on top changes. You can see this in the coloured layers of the platform stack.
Curious what this looks like on your data?
A lakehouse proof of concept shows it in four to six weeks, on your own data.