Selected work
Enterprise & public sector · National digital identity

Biometric liveness for national digital identity

Integrating and maintaining the biometric liveness solution within Digital Dubai / UAE PASS: the mobile SDKs, the servers behind them, and the releases that carried changes into the live service.

  • iOS
  • Android
  • FaceTec
  • Docker Compose
  • MongoDB
  • Python
  • REST APIs
  • Load balancing
Read the case study

The engagement

I was responsible for implementing and maintaining the biometric liveness solution within Digital Dubai / UAE PASS, with ownership spanning technical integration, infrastructure, and coordination with the teams operating the wider platform.

Conceptual view · request path
Mobile app
UAE PASS on iOSdevice SDK
UAE PASS on Androiddevice SDK
Digital Dubai services
Web and REST servicesthe identity journey stays with the platform
Liveness service
FaceTec Server SDKDocker containers, load-balanced
MongoDB clusterhighly available
FaceTec dashboardserver side

The result travels back to the device the same way.

A focused capability

The scope centred on liveness and biometric verification within a much larger identity platform. Maintaining that capability meant working across application and server components, coordinating releases, and keeping implementation guidance clear for the teams integrating and supporting it.

My work included hands-on mobile SDK integration on iOS and Android, deployment planning, testing, knowledge transfer, and ongoing technical support.

Engineering decisions

Paired nodes, separate environments. The server side ran as Docker containers, brought up with Docker Compose on paired, load-balanced application nodes: the FaceTec Server SDK and its dashboard on each. A MongoDB cluster ran on its own database servers for high availability. QA and production were separate environments built the same way, so a release could be proven in one before it reached the other.

Liveness service architecture · redrawn from the deployment guide
UAE PASS appiOSAndroidFaceTec device SDK
Digital Dubai servicesweb and REST, own the identity journey
Load balancerspreads calls across both nodes
Liveness service · production environment
Service APIs · what an integrating team calls
statussession token3D liveness
App node 1 · Docker Compose
  • FaceTec Server SDK
  • FaceTec dashboard
App node 2 · Docker Compose
  • FaceTec Server SDK
  • FaceTec dashboard
MongoDB clusterhighly available, on its own servers
Usage logssent on over HTTPS or SFTP
Python automationworkflows and infrastructure
QA environment · same shapeEvery release is proven here firstthen production

Automated with Python. I wrote the workflows and infrastructure automation around the service in Python, so routine operational work ran from scripts rather than by hand.

One capability, clear boundaries. The app, the user journey, and the identity decisions stayed with the platform. I owned the liveness capability inside it: the server, its dashboard, the deployments, and the guidance the mobile teams built against.

Who ran what
The platform kept
  • The UAE PASS app and user journey
  • Web and REST services
  • Identity and account decisions
  • Security and network requirements
I delivered
  • Liveness server and dashboard
  • SDK integration with the mobile teams
  • QA and production deployments
  • Migration and deployment guides

Three calls for integrators. The service exposed a small, documented surface. A team integrating liveness needed three calls, and the guide covered each one.

Service APIs · names simplified
status
is the service ready
session token
issued before a capture starts
3D liveness
checks the capture, returns the result

Written down for the next team. I wrote the migration and deployment guides end to end, so the teams integrating and operating the service worked from one reference instead of from memory.

The major upgrade

In 2021 the liveness SDK moved to a new major version. The change reached every layer: renamed device APIs on iOS and Android, a new version of the server REST API, and a database schema migration. I planned the upgrade, wrote the procedure, and carried it through QA into production.

SDK major version upgrade · what changed
  1. 01 · DeviceSDK APIs renamed on iOS and Android
  2. 02 · ServerNew REST API version
  3. 03 · DataSchema migrated, backup first
  4. 04 · LicenceProduction licence configured
deployed to QA firstvalidated on iOS and Androidthen production

The order mattered more than any single step. The database was backed up before anything changed, the new version was proven in QA with the mobile teams, and only then did production move.

What I owned

  1. 01

    Hands-on SDK integration with the mobile teams, on iOS and Android.

  2. 02

    Infrastructure and deployment requirements for the liveness solution, across QA and production.

  3. 03

    Workflows and infrastructure automation in Python.

  4. 04

    The major SDK upgrade: device API changes, the new server API version, the database migration, and the rollout.

  5. 05

    Migration and deployment guides, testing support, and knowledge transfer.

  6. 06

    Technical coordination across Digital Dubai stakeholders, delivery teams, and the biometric vendor, with ongoing maintenance.

Where it stands

2environments, QA and production, built the same way
3service APIs documented for integrating teams
1major SDK upgrade carried from QA into production

I maintained the liveness capability from 2021 to 2024, working with government stakeholders, mobile engineering teams, infrastructure and security teams, and the biometric vendor to carry changes into the live service.

It was one capability inside a much larger platform. Owning it well meant owning the details: the application and server components, the releases, and guidance clear enough for other teams to integrate and support it.

WORKING TOGETHER

Need technical ownership of a critical integration? Get in touch.

Let’s talk
NEXT CASE STUDYA shared biometric platform inside a private bank First Abu Dhabi Bank