Fintech
Rebuilding payment rails for scale and trust
3x throughput, 99.97% uptime, 60% fewer failed payments.
Client
NovaPay
Industry
Fintech
Timeline
9 months
Team size
6 people
The challenge
NovaPay's monolithic payment system was failing under peak load, with transaction error rates above 4% during promotional events.
Compliance gaps with PCI-DSS slowed enterprise sales, and the team had no real-time visibility into fraud patterns.
Our approach
We rebuilt the platform on a microservices architecture in Node.js running on AWS, with an event-driven core for transactions.
A real-time fraud detection layer powered by TensorFlow Lite scored every transaction in under 80ms, and the entire system was hardened to PCI-DSS Level 1.
Architecture
[ Clients ] → [ API Gateway ] → [ Microservices ]
↓
[ Event Bus ]
↓
[ Data Layer + Cache ]Outcomes
The results
99.97%
Uptime SLA
3x
Transaction throughput
60%
Fewer failed payments
<80ms
Fraud scoring latency
Technology stack
Node.jsReactPostgreSQLAWS LambdaStripeTensorFlow Lite
Team structure
Tech Lead4× Backend2× FrontendDevOpsQAML Engineer
"CuraLabel didn't just deliver — they made our platform something enterprise customers trust."
Anika Rao
VP Engineering, NovaPay