Paays
A single API that lets auto dealers and lenders verify identity, income, and AML risk in minutes instead of days, turning fraud checks from a paper bottleneck into a real-time decision.
The problem
Auto financing ran on paper and trust.
A dealer would photograph a license, a lender would manually cross-check a pay stub, and by the time identity and income were confirmed, fraud had often already happened or the deal had stalled for two days.
What I built
Four checks, one decision, one API.
Paays combines ID verification, income verification, fraud signals, and AML screening into a single flow a dealer runs once at the desk. I built the system end to end: the biometric/OCR ID pipeline, the income-verification integration layer, the real-time fraud scoring engine, and the dealer/lender API that ties all four together into one decision.
The result isn't four separate reports. It's one clear pass or flag, with the underlying evidence attached, so a finance manager doesn't need to interpret four different vendor dashboards to make a call.
- Identity verification via biometrics + OCR
- Income and employment verification in minutes, not days
- Real-time fraud signal scoring
- FINTRAC-compliant AML, sanctions, and PEP screening
- One API for CRM, desking, and loan origination systems
By the numbers
- $150k
- Saved for one dealer partner in one month
- 2 days → 2 minutes
- For income verification
- 99%
- Reduction in fraud for full-volume partners
- One integration
- Replacing four vendor dashboards
Illustrative figures.
System design
Verify once, decide everywhere.
A dealer-initiated transaction is immediately available to the lender side, so a lender never re-runs a check the dealer already completed. Each verification type runs as an independent, composable check, so a partner can call only what they need without re-architecting their integration.
One verified identity should follow the deal, not get re-checked at every desk it passes through.
Results and trust
Built for regulated decisions.
Built for FINTRAC-regulated AML/KYC requirements, with end-to-end encrypted identity and income data, and designed to sit alongside existing LOS and decisioning engines rather than replace them.
Built with
- Machine Learning
- AI Development
- Computer Vision
- Facial Recognition
- OCR & Document Processing
- Python
- Backend Development
- Full-Stack Web Development
- API Development & Integration
- Database Engineering
- SQL
- FinTech Security
- KYC & AML