Aurora Pay: Real-Time Fraud and Transaction Monitoring

Portfolio project (synthetic data), solution design complete
Cloud-native, event-driven fraud detection platform on AWS, to be provisioned with Terraform
• Completed the solution design for an event-driven fraud detection platform on AWS that screens card, wallet, and
bank-transfer transactions in real time, for a fictional payments company with 3M customers and about 2M
transactions per day. Implementation with Terraform is in progress.
• Documented the end-to-end architecture, event catalog, data model, and architecture decision records (ADRs) for
key trade-offs: latency vs. check completeness, availability vs. fallback risk, and auditability vs. cost.
• Designed to defined targets: p99 card authorization decision under 300 ms, sustained 1,500 TPS with bursts to
5,000 TPS, 99.95% decision-path availability, RPO under 5 minutes, and RTO under 1 hour, with failure analysis
for each major component.
• Designed security and compliance controls aligned with PCI DSS and PIPEDA: least-privilege access, encryption in
transit and at rest, tokenized card data, and immutable 7-year audit retention with tiered storage.
Tech Stack: AWS, Terraform, GitHub Actions, Event-Driven Architecture, Solution Architecture, Architecture Decision Records