Cloud-Based IoT Solution for Smart Aroma Diffusers
A detailed knowledge-base record capturing the full history, architecture, value, and lessons learned from a multi-year IoT product engagement.
Foreword
The Client approached us with a vision to bring a connected, app-driven experience to the world of premium aroma diffusion. Rather than building yet another smart appliance, our partnership centered on shaping an end-to-end IoT product — from cloud architecture and a native mobile app down to the firmware running on the device itself.
Through a long-running dedicated-team engagement, we helped the Client de-risk their hardware roadmap, design a scalable cloud platform, and step in as the firmware development partner when an external vendor could not deliver. The result is a coherent, production-ready ecosystem that links physical devices, a customer mobile app, an admin dashboard, and the Client's e-commerce flow.
1. Project Summary
| Field | Value |
|---|---|
| Project Name | ScentCloud (project codename) |
| Industry | IoT / Connected Consumer Devices |
| Purpose / Business Objective | Development of a cloud-managed, app-controlled aromatic oil diffuser ecosystem. |
| Status | Pre-Production |
| Project Source | Referral |
2. Client Description
| Field | Value |
|---|---|
| Client | A US-based scent marketing and aroma diffusion company (under NDA). |
| Company Size | Approximately 200–500 employees. |
| Engagement Type | Dedicated Team — Time & Materials. |
| Business Vertical | IoT, Smart Home, Scent Marketing & Retail. |
| Geo Coverage | United States. |
| Annual Revenue | Mid-sized (under NDA). |
3. Story
The Client is a North American consumer-tech and scent-marketing company offering premium aroma diffusion solutions to residential customers and commercial venues. Their growth strategy depended on turning a traditionally analog product — the scent diffuser — into a connected device that could be managed remotely, personalized per environment, and integrated with their online store.
The Client approached us looking for a long-term technology partner that could cover the full IoT stack: cloud services, mobile applications, and (eventually) device firmware. A hardware manufacturer had already been pre-selected, and the diffuser model was meant to communicate over MQTT. Our role started with a discovery phase based on the existing technical documentation and a single test device, and grew into ownership of the cloud platform, the mobile app, the DevOps pipeline, and ultimately the firmware itself.
Today, the engagement is focused on stabilizing the platform, finalizing firmware for a production-ready hardware revision, and preparing the ecosystem for a commercial launch under a strict release schedule.
4. Requirements & Challenges
When the engagement started, the Client had a clear product vision but limited in-house engineering capacity to deliver a connected device end-to-end. Bringing the product to life required coordinating cloud, mobile, hardware, and firmware tracks in parallel, while staying flexible enough to absorb a mid-project hardware vendor change. Key requirements and challenges included:
- Replacing manual device management with a centralized cloud system — Without a centralized platform, configuring and monitoring the diffusers required manual effort, leading to operational overhead and human error. The Client needed a single cloud system to provision, configure, and observe devices in the field.
- Eliminating manual firmware updates — Each device required firmware updates to be applied manually, which meant inconsistent versions across the fleet and additional maintenance time. An over-the-air update mechanism was a critical requirement before scaling.
- Reducing customer support load through automated diagnostics — Because there were no automated diagnostics or self-service configuration tools, even routine device setup generated support tickets. A better customer-facing configuration flow and remote troubleshooting tooling were needed.
- Connecting e-commerce and IoT provisioning — Online sales were disconnected from device activation: customers could buy a diffuser, but post-sale provisioning, oil reorder flows, and usage analytics had to be assembled separately. The platform needed to close that gap.
- Absorbing a hardware vendor change without slipping the roadmap — Mid-project, the original hardware vendor could not meet the firmware requirements. A new US-based hardware partner was engaged, and we had to take ownership of firmware development to keep the schedule realistic.
5. Solution Overview
Phase 1 — Cloud-Based IoT Platform Concept and Initial Implementation
The Client approached us with a request to develop a cloud-based IoT solution for managing aromatic oil diffusers. At the initial stage, the business idea was to let end customers remotely control their diffusers, monitor oil levels, and reorder oils through a mobile application connected to the device. A hardware manufacturer had already been selected, and the diffuser model supported communication via an MQTT broker. Based on the provided documentation and a test device, our team conducted a discovery phase and proposed a full Cloud-Based Solution architecture: cloud infrastructure, API, MQTT broker, dashboard, and mobile app.
Phase 2 — Firmware Development and Hardware Vendor Transition
During cooperation with the initial device manufacturer, it became clear that the vendor could not fulfill all technical requirements — particularly the ability to modify the firmware. A new US-based hardware vendor was engaged to design a custom device from scratch, covering electronics, firmware, and industrial design. However, delays and capability gaps prompted us to take over firmware development responsibilities. The team expanded with an embedded developer and successfully continued the implementation. The project is currently at the pre-release stage.
Deliverables
- Software Development — Full-stack development of the cloud platform and a native mobile application.
- Firmware Development — End-to-end firmware creation from scratch.
- Quality Assurance (QA) — Functional, integration, and performance testing.
- CI/CD and DevOps — Automated pipelines, GCP infrastructure configuration, and environment management.
6. Features
The ScentCloud platform is composed of several tightly integrated components. Together they cover the full lifecycle of a connected diffuser — from manufacturing and provisioning, through everyday consumer use, to fleet-level analytics and marketing operations.
Cloud Platform
A scalable, microservice-based backend hosted on Google Cloud Platform and orchestrated with Kubernetes. The platform exposes secure APIs for the mobile and admin clients, brokers MQTT traffic to/from devices, and runs asynchronous workflows on top of RabbitMQ. Postgres and Redis handle transactional and cache data respectively.
Native Mobile Application
A consumer-facing mobile app built for iOS and Android in React Native. End users can pair their diffusers, schedule scent programs, monitor oil levels in real time, receive low-oil notifications, and reorder oils directly through an integrated checkout flow. Real-time device state is synchronized over Socket.IO.
Admin Dashboard
A React-based dashboard for the Client's operations team. It surfaces fleet health, device telemetry, firmware versions, and per-customer activity. Admins can issue remote commands, manage release cohorts, and inspect troubleshooting data without touching the underlying database.
Custom Firmware
Firmware developed in-house from scratch and tailored to the new US-based hardware design. It covers MQTT connectivity, secure provisioning, device diagnostics, and the runtime that drives the diffusion logic. The firmware is shipped through the platform's OTA pipeline.
OTA & Release Management
An over-the-air update and release management subsystem allows new firmware versions to be deployed gradually across cohorts of devices. It supports staged rollouts, rollback, and per-version compatibility checks, enabling safe production updates without on-site service.
Telemetry & Usage Statistics
A telemetry and analytics subsystem captures detailed device-usage data — sessions, scents used, oil consumption, error events. The data feeds operational dashboards today and is structured to support future marketing and product analytics use cases.
E-commerce & Payments Integration
Stripe is integrated for in-app purchases and oil reorders, linking the IoT layer back to the Client's commercial flow. Customers can move from a low-oil notification to a completed order without leaving the app.
7. Stack
| Layer | Technology |
|---|---|
| Frontend | React, Redux, Socket.IO |
| Mobile | React Native |
| Backend | Node.js, NestJS, Socket.IO |
| Database | PostgreSQL, Redis |
| Messaging / Events | RabbitMQ, MQTT |
| DevOps | GCP, Kubernetes, Docker, CI/CD |
| Integrations | Stripe |
| Firmware | Embedded C/C++ (custom firmware stack) |
Key Architectural Decisions and Justification
| Decision Area | Description & Justification |
|---|---|
| Microservice Architecture | Event-driven architecture based on RabbitMQ ensures scalable communication between microservices and supports asynchronous processing of IoT events. |
| OTA & Release Management | An over-the-air firmware update and release management system enables remote version control and coordinated rollouts. |
| Usage Statistics System | A telemetry and analytics subsystem collects detailed data on device usage, planned for future marketing analysis. |
8. Engain AI-Native Approach
ScentCloud is built and supported using Engain's AI-native delivery model. Senior engineering expertise is combined with AI agent orchestration so that 20% of effort goes into core development and 80% into automated, AI-driven maintenance — radically cutting operational cost without giving up output quality.
| Capability | How it applies to ScentCloud |
|---|---|
| AI agent orchestration | AI agents are wired into the delivery pipeline to assist with code generation, review, and triage across the cloud, mobile, and firmware tracks. |
| AI-augmented engineering workflows | Senior engineers use AI-augmented workflows for faster delivery, fewer bugs, and lower cost compared with traditional agency staffing. |
| AI-powered QA | Functional, integration, and regression tests are accelerated by AI-assisted test generation and anomaly detection on top of the existing QA practice. |
| Clickable prototype on kickoff | Prototype-first delivery (clickable prototype within 24 hours of kickoff) was used to validate UX and device-control flows before full implementation. |
| AI-automated maintenance | Once features ship, AI agents take over routine maintenance, bug triage, and auto-fix patterns — targeting an ~80% reduction in ongoing support effort. |
| Automated monitoring & observability | Full observability from day one: performance metrics, anomaly detection, and proactive optimization across the GCP/Kubernetes platform and the device fleet. |
| Strategic AI partnerships | Direct collaboration with leading AI providers ensures the most capable models are used for each specific task in the pipeline (codegen, review, monitoring). |
| Industry coverage | Engain delivers the highest ROI in high-volume sectors — Real Estate, E-commerce (UK & US), Legal & Security, and Service Sector. ScentCloud's consumer e-commerce and oil-reorder flows fit directly into this focus. |
9. Project Timeline
| Phase | Period | Key Milestones |
|---|---|---|
| Discovery | Apr 2025 – May 2025 | Study of existing technical documentation and Client requirements. |
| Architecture | Jun 2025 – Jul 2025 | Design of the cloud-based IoT architecture against Client requirements. |
| Development | Aug 2025 – Present | Implementation of cloud services, mobile app, and firmware; continuous QA. |
| Release Preparation | Mar 2026 – Present | Final pre-release activities and production environment setup. |
10. Core Delivered Features
| Feature | Description & Business Benefit |
|---|---|
| Discovery Phase | Identified potential risks related to the Client's initially selected hardware vendor, helping to avoid delays and technical pitfalls. |
| Cloud-Based Solution | Built a scalable IoT cloud infrastructure on GCP and Kubernetes, following best practices for security and maintainability. |
| Mobile Application | Developed a native mobile application from scratch for remote control, monitoring, and management of diffusers. |
| Firmware Development | Took over full firmware development to accelerate delivery and ensure compatibility between hardware and cloud components. |
| OTA Update Pipeline | Implemented a release-management and over-the-air update mechanism to keep the device fleet on supported firmware versions. |
| Usage Telemetry | Designed a telemetry layer that supports operational monitoring today and marketing analytics in the next phase. |
11. Project Outcomes
As the engagement progressed, the platform matured into a coherent, production-ready IoT ecosystem. Beyond writing code, our team consistently challenged risky technical decisions and absorbed scope that was originally meant to sit with external vendors. The work has resulted in the following outcomes:
- Prevented costly hardware design flaws — Our engineering and firmware expertise helped the Client identify and avoid a critical hardware architecture issue that could have led to a recall of the first production batch and significant release delays.
- Single accountable partner across cloud, mobile, and firmware — By stepping into firmware development when the third-party vendor faltered, we consolidated cloud, mobile, and device-side responsibilities under one delivery team — reducing coordination overhead and shortening feedback loops.
- Production-grade cloud foundation — The platform runs on a scalable GCP/Kubernetes foundation with CI/CD, containerization, and event-driven messaging via RabbitMQ. It is ready to support the device fleet at launch volumes and beyond.
- Safe, controllable rollouts — An OTA and release-management subsystem allows the Client to ship firmware updates gradually across cohorts, with rollback options. This minimizes the risk of fleet-wide regressions once the product is in customers' hands.
- Foundation for marketing and product analytics — The telemetry subsystem captures structured usage data that the Client can use to drive marketing decisions, refine the consumable (oil) reorder flow, and prioritize the post-launch product roadmap.
Project Information Record | ScentCloud | Confidential