← Back to projects
GyL Group · In production

AI-Powered Biometric Validation

Complete solution (backend + frontend) for identity validation through facial recognition, with liveness detection and immutable traceability, integrated with RENAPER.

500+
Daily verifications in production
99.8%
Biometric validation precision
Liveness
Motion detection

The Problem

A government agency needed to validate people's identity in real time, comparing against RENAPER data, with a liveness check to prevent fraud and traceability that makes every validation auditable.

  • • Real-time identity validation against RENAPER data
  • • Liveness detection with motion capture to prevent impersonation
  • • Full traceability of every validation (who, when, with what result)
  • • Demanding precision for a government-scale system

My Process

1. Facial recognition with AWS Rekognition

Implemented identity validation through facial recognition, integrating AWS Rekognition for biometric matching.

  • ✓ Facial comparison against RENAPER data
  • ✓ Liveness check with motion detection
  • ✓ 99.8% precision validated in production

2. Node.js/NestJS backend integrated with RENAPER

Built the RESTful API in Node.js/NestJS that consumes RENAPER services, overcoming government documentation gaps.

  • ✓ RESTful API consuming RENAPER services
  • ✓ Overcame the agency's documentation limitations
  • ✓ Angular frontend for biometric capture

3. Immutable traceability with AWS QLDB

Implemented immutable storage with AWS QLDB so every validation is recorded and auditable.

  • ✓ Immutable ledger with AWS QLDB
  • ✓ Full traceability of validations
  • ✓ Auditable history for the agency

4. Scale in production

The system runs in production processing 500+ daily verifications reliably.

  • ✓ 500+ daily verifications in production
  • ✓ Stable operation at government scale
  • ✓ 99.8% precision validated on real traffic

The Solution

Real-time identity validation

The person is validated against RENAPER via facial recognition, with a liveness check that rules out impersonation.

• Facial recognition with AWS Rekognition
• Liveness check with motion detection
• Direct integration with RENAPER

Immutable traceability

Every validation is recorded in an immutable ledger, making the process auditable end to end.

• Immutable storage on AWS QLDB
• Complete record of every validation
• Guaranteed auditability for the agency

Precision validated in production

The system operates at real scale with verifiable metrics.

• 99.8% precision on real traffic
• 500+ daily verifications
• Stack: Node.js/NestJS + Angular + AWS (Rekognition, QLDB)

Results & Impact

99.8%
Biometric precision
500+
Daily verifications
100%
Validations with traceability
Prod
Running since 2024

Key Learnings

  • ✓ AI to production: liveness detection with AWS Rekognition showed that taking AI to production requires integrating vision, backend and traceability, not just a model.
  • ✓ Agency integration: working with RENAPER meant resolving incomplete documentation through direct requirements gathering and iterative testing.
  • ✓ Ledger as a guarantee: immutable traceability (QLDB) is what made the system auditable and defensible before the agency.

Tech Stack

Backend

• Node.js
• NestJS
• TypeScript
• PostgreSQL

Frontend

• Angular
• Biometric capture

Cloud & AI

• AWS Rekognition
• AWS QLDB
• AWS S3
• AWS Lambda

Need identity validation or AI in production?

If your system needs facial recognition, liveness detection or immutable traceability, let's talk. I did it at government scale.