Real-time Fraud Detection
Machine learning models for fraudulent transactions
- Sector
- Financial Services
- Year
- 2026
- Duration
- 7 months
- Status
- In Development
- Client
- Client project
About the project
Development of a machine learning system to detect fraudulent transactions in real time. Every operation is scored as it happens and, above a given risk threshold, is routed for human review before being authorised.
Technologies
Python
Machine Learning
REST API
PostgreSQL
Features
- Real-time scoring
- Review queue
- Audit log
Results
Risk scoring applied to every transaction, with human review on borderline cases
