Selected work
AI case studies: built, shipped, still running.
Engagements from the last few years, with what changed after they went live.
01
AI-Driven Fraud Detection for FinTech
CapitalEdge was losing $6M annually to payment fraud.
FinTechGNNReal-time MLRisk
CapitalEdge
FinTech
62% reduction in fraud losses · 99.1% detection precision · 12ms average inference latency
02
Predictive Maintenance for Global Manufacturing
NexaCorp's global facilities were struggling with ageing legacy systems producing millions of data points but zero actionable insights.
ManufacturingMLIoTPredictive Analytics
NexaCorp
Manufacturing
30% reduction in downtime · $1.8M annual savings · 94% anomaly detection accuracy