Six hours to one, at global scale

The challenge

When I stepped in as Director, the reporting-and-billing backend was a bottleneck. Processing windows were long and accuracy issues surfaced across global markets — exactly the kind of problems that erode trust in the numbers customers are billed on.

What I did

The call was to rebuild rather than patch — the failures were architectural, and patch cycles were consuming more capacity than a rebuild would. On that basis I led a redesign of the reporting-and-billing backend — rethinking how data moved through ETL, normalization, and aggregation so the system could scale with the business instead of fighting it.

This was a system of record that couldn't go down: clients are billed on these numbers, and there was no maintenance window long enough to swap it out. So the cutover was engineered, not hoped for — the rebuilt pipeline ran in parallel with the old one, and we reconciled their outputs against each other until the new system had earned the switch. Markets cut over with zero downtime, and the old system stayed as the safety net until the numbers proved it wasn't needed.

The outcome

The redesign cut the daily processing window from six hours to one — at the same data volume — and cut the data-error rate roughly 4× across all markets. Client finance teams felt it directly: monthly reconciliation closed in half the time. In the same chapter I launched a net-new master data management platform (0-to-1) that turned standalone tools into a connected ecosystem — today the redesigned pipeline moves a portfolio spanning roughly 2.4 trillion rows.

What it proves

Technical execution at scale, a bias for reliability and data quality, and platform thinking — turning a bottleneck into foundational infrastructure the business could build on.