Menu
Ello
Industry : Artificial Intelligence | Digital Media | Personal
Tech Stack : AWS Serverless | Node.js | TypeScript | DynamoDB | AWS IVS | AWS Rekognition | AWS IoT Core | Firebase | SES
The ELLO team came to us with a clear and ambitious vision: a smart home security platform that gives families real intelligence about their home, not just footage. A system that knows the difference between a family member walking through the door and a stranger stepping onto the property. One that streams live video with no meaningful lag, sends alerts that actually mean something, and manages the cameras remotely without requiring any technical knowledge from the people using it.
What they needed from us was the backbone that would make all of that possible. Zimozi took full ownership of the backend for ELLO’s MVP, designing and building the entire server-side architecture from the ground up, covering real-time video infrastructure, AI-powered detection, IoT device management, notification systems, and the security layer that holds it all together.
A smart home security backend is not a simple product. It sits at the intersection of real-time performance, computer vision, IoT reliability, and the kind of security that families genuinely depend on. Here is what we were working to solve.
Families checking on their home from across the city, or across the world, need a feed that is immediate and reliable. Any delay in the stream makes the product feel untrustworthy, and that is not a feeling you recover from easily.
A home security system that cries wolf loses its audience quickly. The backend needed to distinguish meaningful events from everyday household activity, and to recognise the family so their movements never triggered a false alarm.
Configuring cameras, pushing firmware updates, setting schedules: all of this needed to happen remotely, over a secure channel, in a way that felt effortless to the end user.
This is a platform that handles home surveillance data for real families. End-to-end security, proper access controls, and a clean audit trail were not nice-to-haves, they were the foundation everything else had to sit on.
Our development team designed and deployed FabriceAI, a robust AI-powered web application tailored to present Fabrice’s viewpoints through a seamless, engaging user interface.
We built ELLO’s backend as a fully serverless AWS architecture, with over 100 Lambda functions and more than 20 DynamoDB tables, deployed on Singapore infrastructure and built to scale cleanly as the user base grows.
CI/CD automation and end-to-end JWT-based security were baked in from the very beginning, not bolted on afterwards.
For live video, we integrated AWS IVS to deliver low-latency streams that reach the mobile app in near real time. The AI detection layer runs through AWS Rekognition, which watches each feed for motion, intrusions, and fall events, and importantly, recognises registered family members so their presence never generates a false alert. IoT device management runs over a secure MQTT channel through AWS IoT Core, giving the team full remote control over camera configuration, firmware, and scheduling.
When an event is detected, AWS SES and Firebase deliver notifications instantly across email, push, and in-app channels simultaneously.
AWS IVS integration delivers low-latency live feeds to the mobile app, giving families reliable, high-quality access to their home cameras from anywhere.
AWS Rekognition analyses each video feed to detect motion, intrusions, and falls, while recognising registered family members to keep false positives to a minimum.
A secure MQTT channel via AWS IoT Core handles remote camera configuration, firmware updates, and scheduling, all manageable from the app without any technical expertise.
Detection events trigger instant alerts across push notifications, email, and in-app channels simultaneously through Firebase and AWS SES.
JWT-based authentication, role-based access controls, and a microservices architecture designed to keep family surveillance data secure at every layer.
100+ Lambda functions and 20+ DynamoDB tables on Singapore infrastructure, with CI/CD automation and a clean architecture that scales with the product without a rebuild.
ELLO launched its MVP with a backend that was production-grade from day one. The architecture handles real-time video, AI detection, IoT device management, and instant notifications without any of the instability that typically comes with a first version. The ELLO team could go to market with confidence, knowing the foundation underneath them was built to grow.
