Work

Authentify Fraud Detection

UI lead for fraud detection system at BNY Mellon

The Problem Space

Financial transactions require robust fraud detection. The challenge was building a system that could identify fraudulent business transactions based on user behavior patterns, while remaining usable for investigation teams.

Working Conditions

  • Accuracy: False positives create operational burden; false negatives create risk
  • Complexity: Multiple data signals needed to be presented coherently
  • Urgency: Investigation teams needed quick access to relevant information

Architectural Choices

Designed User Experience with prototyping tools, securing business approval before implementation.

Built UI optimized for investigation workflows, not just data display.

Dockerized independent workflow engine to BNY Mellon's Extreme Cloud platform for scalability.

Results

System enabled identification of fraudulent transactions based on user behavior.

UX design process established prototype-first approach that influenced subsequent projects.