Mastercard was experiencing intermittent failures in the clusters responsible for routing fraud scoring on payment transactions. When these clusters degraded, the result was point-of-sale delays and declined transactions — a problem affecting billions in transaction volume annually. I was brought in to lead a $250K paid proof of concept to determine whether failures could be predicted and traffic rerouted before they caused customer-facing disruption.
We trained anomaly detection models on Splunk telemetry data, identifying cluster health signatures that preceded failures before they materialized. The models powered an automated routing layer that redirected transaction traffic proactively — eliminating the latency and declines caused by reactive failover.
After a successful proof, I identified the opportunity to replicate the architecture across six other Mastercard services. I built the stakeholder relationships needed to position that expansion, then structured the deal to protect it. The original contract included a termination-for-convenience clause that left us exposed across a multi-year buildout. I renegotiated: any early exit would require Mastercard to pay more than 60% of remaining contract value — converting meaningful commercial risk into downside protection.
Six systems built across Mastercard services over three years. Billions in annual transaction volume protected from cluster-failure-driven disruption. Contract restructured so more than 60% of remaining value was due on early termination.