Consortium payments security

This is a corporate project. Due to confidentiality agreements, source code, real data, and sensitive details cannot be shared. The content below focuses on technical challenges, architectural decisions, and key learnings from the project.

Context

The fintech managed dozens of groups with thousands of billets every day. Any delay in reconciliation blocked payouts and created room for insider fraud.

Problem

The legacy pipeline was synchronous and tied to the banking core. During peaks billets piled up, forcing the team to release each payment manually with barely any traceability.

My role

I partnered with the risk squad to map every touchpoint, designed the event model and built an anti-fraud layer that tracked each billet from registration to the investor payout.

Technical challenges

• Keep eventual consistency across multiple banks and gateways.
• Produce audit trails even with thousands of events per minute.
• Recalculate payouts when a billet was reversed or flagged as suspicious.

Decisions and solution

I implemented a Node.js orchestrator with SQS queues and DLQs per operation. Anti-fraud checks called internal and external services in parallel while appending the status to the same event. Critical data landed in a partitioned PostgreSQL plus real-time streams for the finance team.

Result/Impact

Release SLA dropped below five minutes even during campaigns with 3x more billets. Fraud attempts surfaced 40% faster thanks to automatic alerts routed to the SOC.

Stack

• Node.js + TypeScript
• AWS SQS and Lambda
• Partitioned PostgreSQL
• Grafana + Loki for observability