🎯Solution Overview

This transformative solution is built on a collaboration between EBS and Tech5, where the Deduplication Engine is composed of two primary layers:

Demographics (EBS)

EBS handled the core Deduplication Engine and the Demographics layer:

• Event-Driven Architecture: Utilizes Kafka for high-speed ingestion and Celery for orchestrated background processing.

• Scalability: Features KEDA-driven auto-scaling to handle massive data batches and traffic spikes.

• Demographics Matching: Implements Elasticsearch for fuzzy demographic matching to identify potential duplicates based on textual data.

Biometrics (Tech 5)

Tech5 handled the Biometrics layer, providing specialized technologies for identity verification:

• OMNIMATCH ABIS: A powerful 1:N deduplication engine for facial identification.

• Passive Liveness Detection (T5-LDS): A "fail-fast" verification system that stops spoofing attempts (e.g., photos of photos) before expensive processing occurs.

These components are integrated through the EBS-managed Deduplication Engine, ensuring a seamless flow from ingestion to adjudication.

Key Deliverables

The EBS-TECH 5 Deduplication Solution delivers:

• A "Golden Record": Establishes a clean, centralized subscriber database that serves as the single source of truth.

• Instant Duplicate Prevention: Real-time biometric verification during customer onboarding processes prevents bad data from entering the system.

• Automated Adjudication: Programmatically isolates duplicate and fraudulent records into a manual review queue with a full evidence package.

• Regulatory Compliance: Ensures full alignment with NCC mandates for a verifiable "one-person-one-ID" database.

Strategic Impact

This project showcases MTN's forward-thinking strategy in data governance. By automating the detection of identity fraud—such as one person registering multiple SIMs with different names but the same face—it sets a new benchmark for telecom customer management in Nigeria, driving growth and operational efficiency.