Back to projects
ANTIC
AI fraud detection & SMS classification system with real-time alerts.
Security · Real-time classification · Backend systems
Security / AI system
Screenshots
What it solves
Users and organizations needed a way to detect fraudulent messages and classify SMS content in real-time, with secure access and actionable alerts.
Outcome / impact
Shipped backend services for real-time fraud detection workflows with secure access and alert-driven responses.
My role
- Backend APIs
- System integration
- Authentication design
Key technical challenge
Integrating ML inference into a secure, low-latency API pipeline while keeping authentication and alerting reliable.
Highlights
- Secure RBAC + 2FA flows for admin access
- ML inference integration for classification and alerting
- Low-latency API pipeline for real-time detection workflows
Architecture (high level)
- Client-server architecture: React admin UI + Laravel core API
- Python (Flask) microservice for ML inference integrated via secure APIs
- Event/queue-ready alerting pipeline (email/notifications)
Technologies used
Python (Flask), Laravel, PHP, React, AI model integration, Auth, APIs