Logistics & Operations · Social Housing

Australia's Homeless Housing Platform

A cloud native system where housing providers manage homes, rooms and where social services place clients for accommodation in real time

10 weeks

Focus: SAAS, DATA-PLATFORMS, THIRD-PARTY-INTEGRATION, CUSTOM-WEB-APP
Delivered: 11 Jun 2025
Location: Sydney, Australia

Outcome

65% faster placements, 40% better bed-occupancy accuracy in the NSW pilot; live in ten weeks.

Project Background

Australia’s homelessness crisis has grown more complex in recent years, with emergency beds often under-utilised simply because agencies, shelter operators and outreach teams lacked a common digital channel. The client, a social-impact start-up based in New South Wales, envisioned an online hub that would let homeless individuals, support workers and accommodation providers collaborate through a single secure workflow.

Their objective was twofold: improve placement speed for people in need and create a financially sustainable SaaS product that could scale interstate. Regulatory obligations added another layer of urgency: the platform had to satisfy GDPR, ISO 27001 and local Essential 8 controls from day one. Robust Devs was invited to turn this vision into an MVP within ten weeks, laying the technical foundation for a nationwide roll-out while ensuring every byte of personal data stayed onshore and encrypted in transit and at rest.

The challenge

From the outset, the project faced several intertwined obstacles. First, three distinct user journeys homeless client, support worker and housing provider needed vastly different interfaces yet had to share a single codebase for maintainability.

Second, sensitive identity documents and chat transcripts would traverse the system daily, demanding airtight encryption, role-based permission gates and an auditable moderation trail.

Third, the service could not charge its most vulnerable users, so the business model relied on precise subscription billing for providers and automated per-bed fees, each synced with fluctuating room occupancy data.

Fourth, shelters span urban and remote regions; the discovery engine therefore required geospatial search plus nuanced filters for age, cultural background and support needs, all delivered with sub-second response times on low-power mobile devices.

Finally, every compliance checkbox had to work from launch, from GDPR portability requests to IRAP logging to Essential 8 patch cadence, not as later add-ons, lest trust evaporate among government stakeholders and advocacy groups.

How we solved it

Robust Devs crafted a cloud-native stack that balanced velocity with governance. The front end combined React 18, TypeScript and Next JS for statically generated marketing pages and dynamically rendered dashboards, giving each role a tailored workspace while sharing a single component library.

On the server side, Node JS with Nest JS orchestrated GraphQL APIs backed by a partitioned PostgreSQL cluster on Azure. Auth0 provided OAuth2 flows, then a fine-grained access-control layer enforced JSON-Web-Token scopes across micro-services. Personally identifiable information was AES-256 encrypted before persisting; file uploads landed in Azure Blob Storage behind private endpoints, while chat content streamed through a WebSocket gateway that fed an abuse-detection queue in Python.

Stripe’s metered-billing APIs automated monthly subscriptions and per-bed usage fees, and GitHub CI pipelines wrapped the whole codebase in Snyk scans and OWASP ZAP regression tests. Finally, Terraform scripts codified the entire environment, enabling one-click deployment to an isolated Australian region so auditors could trace every change from commit to container.

A five-layer stack. At the top, three tailored workspaces: one for a person seeking a bed, one for a support worker placing clients, one for an accommodation provider managing rooms. Beneath them a single shared component library keeps three very different journeys in one maintainable codebase. Below that a GraphQL API checks JSON Web Token scopes per field, personal data is encrypted with AES-256 before it is written, and the whole environment runs in a single Australian region so no personal data leaves the country.

Map-Based Housing Discovery

At the heart of the platform sits an interactive map that visualises every registered bed in real time. Built with Mapbox GL and server-side tile caching, it allows users to pan effortlessly across city blocks or zoom out to statewide view without data drag. Adaptive filters let a youth search only age-appropriate shelters, while cultural-support toggles highlight providers versed in language or community-specific care. Availability indicators update every minute via a lightweight WebSocket channel, turning coloured pins from green to amber the instant a bed is reserved, eliminating the frustration of phoning full shelters. Administrators can geofence sensitive listings so, for example, women’s refuges appear only to verified social workers. The map also anchors contextual actions: homeless users can trigger guided applications, support workers can shortlist options for multiple clients and providers can edit vacancies directly from the pin pop-up. By merging geospatial clarity with nuanced access rules, this feature reduces placement time from hours to minutes.

Technologies

What we built it with

Front-end: JavaScript
ReactJS
HTML5
CSS3
Back-end: PHP
Laravel
Python
DevOps: Gitlab CI
Digital Ocean
Database: MySQL
Integrations: Stripe
Auth0
Mapbox

What shipped

The MVP went into production ten weeks after the engagement started. Support workers got a real-time map of available beds, automated provider verification, and a command centre for coordinating placements, with resident data encrypted end to end and every change written to an immutable audit log. Early pilots in New South Wales recorded a 65 percent reduction in average placement time and a 40 percent increase in bed-occupancy accuracy.

Someone without a home can find a safe option, read reviews of it, and reach a support worker from a phone in minutes. The Terraform playbook and the modular services behind it mean interstate expansion is largely a configuration exercise rather than a rebuild.

How we built it

Shared one component library across three very different roles, and encrypted identity data before it was ever stored. A homeless client, a support worker and a housing provider need different interfaces, but three codebases would not have shipped in ten weeks. React 18, TypeScript and Next.js gave each role its own workspace on one shared library, with Nest.js GraphQL APIs on a partitioned PostgreSQL cluster in Azure behind them. Because identity documents and chat cross the system daily, personal data is AES-256 encrypted before it persists, files sit in Azure Blob Storage behind private endpoints, and chat streams through a WebSocket gateway into an abuse-detection queue. Stripe metered billing charges providers per bed, so the service stays free for the people it exists for.

In the client's words
“Our idea has been turned into a user-friendly, safe platform by the Robust team, which at last allows us to concentrate on housing people rather than chasing spreadsheets.”
NDA - Signed

More case studies

View all work

Have a project like this in mind?

Tell us what you're building. We'll show you how we'd approach it and what it would take to ship.

Schedule a meeting See all case studies