SaaS
2025

NUU

AI-powered property matching platform for the Australian rental market.

NUU

Hero Image

Main project showcase

1920 × 1080

Client

NUU Systems

Industry

PropTech / Real Estate

Timeline

8 weeks

Year

2025

Services

Brand Strategy · UI/UX Design · Design System · Frontend Development · Backend Development · AI/ML Integration · Database Architecture

01 — The Challenge

Understanding the problem

The Australian rental market is notoriously competitive and frustrating. Renters struggle with fragmented listings across multiple platforms, impersonal search filters that don't capture lifestyle preferences, and an overwhelming number of unsuitable results. Traditional property search interfaces require users to think in technical terms rather than expressing what they actually want: 'a beachy vibe near good coffee shops' or 'somewhere quiet for my family near trains to the CBD'.

NUU

Gallery

Challenge context visualization

1600 × 900

02 — The Approach

How we tackled it

We took an AI-first design approach, reimagining property search as a conversation rather than a form. We implemented GPT-4o as a 'property concierge' that extracts structured preferences from natural language. We developed a sophisticated 7-factor weighted scoring algorithm that evaluates location, budget, features, amenities, transport, property type, and bedrooms. We created an industrial 'operating system' aesthetic that positions NUU as a premium, tech-forward solution.

NUU

Gallery

Design process & iterations

1600 × 900

NUU

Gallery

Implementation details

1600 × 900

03 — The Solution

What we built

NUU is a full-stack property matching platform featuring a conversational AI interface powered by GPT-4o that guides users through preference discovery in 2-3 natural exchanges. The multi-factor matching algorithm scores properties across 7 dimensions with bonuses for exceptional matches. The frontend features an immersive 3D architectural visualization built with Three.js, scroll-triggered animations, and a dark industrial design language. The backend runs on Vercel serverless with Supabase PostgreSQL, including pgvector for semantic search and PostGIS for geospatial queries.

Tech Stack

React 19ViteTypeScriptThree.jsReact Three FiberOpenAI GPT-4oSupabasePostgreSQLpgvectorPostGISVercelTailwind CSS

NUU

Hero Image

Full showcase view

1920 × 1080

04 — The Outcome

Results & impact

The platform transforms the rental search experience from a tedious filtering exercise into an engaging conversation. The AI accurately extracts preferences from casual language and the matching algorithm surfaces relevant properties. The industrial 'operating system' branding differentiates NUU in a market full of generic real estate interfaces. The architecture is built for scale with vector search ready for real-time property data feeds.

2-3

Conversations

7

Scoring Factors

99%

Match Precision

<2s

Response Time

NUU bypassed the traditional rental friction. My application was processed and approved in 48 hours. Efficiency at its finest.
BB

Batbold B.

Engineer, Sydney