How the Team Works

Every payout and every exceptional deposit credit requires independent approvals.

Changes to custody, identity and the ledger are reviewed before they ship and remain covered by tests after they ship.

Operations run to written procedure: one node at a time, the follower first, and a release identified by its checksum.

An incident is written down with what happened, what changed and what prevents its return.

Duties are covered by defined roles, so every task has a clear owner.

Infrastructure and Availability

The platform runs on redundant multi-node infrastructure, with the application and its database each carried by more than one node.

Traffic is served through a global edge network to an application that runs as an active-active pair of nodes with automatic leader failover.

Writes are accepted by the node holding leadership while the second node carries the same state, so the loss of a node reduces capacity and the record stays whole.

The database streams its write-ahead log continuously to a replica on a separate node.

Identity evidence is held in object storage separate from the application database.

What the Team Is Measured On

Four measures decide whether this platform is being run well.

• Whether customer funds reconcile against on-chain balances.
• Whether every payout passed through the multi-approval pipeline, every time.
• Whether a defect that reached a user can reach another one.
• Whether a user can trace what happened to their own balance from their own record.

Growth follows controls.