Biometric Identity Platform
A distributed Automated Fingerprint Identification System for a national identity program — mTLS edge agents, GPU-accelerated matching, and multi-modal biometric verification at scale across remote enrollment sites.
The Challenge
A national identity program needed to enroll and de-duplicate millions of citizens across geographically remote sites — with intermittent connectivity, strict data-protection requirements, and identification times that had to stay usable at population scale.
Our Approach
We built a distributed AFIS with mTLS edge agents that enroll and match locally, then reconcile with the central system. Fingerprint matching runs on GPU via GoMLX and a binary MCC algorithm, cutting identification time by orders of magnitude. Templates are protected in transit and at rest, and the pipeline extends to face and document modalities.
The Outcome
A production system performing biometric de-duplication and identification at national scale — with edge resilience for low-connectivity sites and matching fast enough for real-time enrollment.
More from AI & Biometrics
Related work
Need something similar?
Tell us about your project and we'll share how we'd approach it.