Skip to content
← All work
AIClient work

Luxconn

Satellite-imagery AI that turns high-value properties into scored contractor leads

Role
Lead full-stack engineer (web v1 and v2 solo; mobile app with one collaborator)
Client / product
Luxconn
Years
2025
Status
Archived

Overview

Luxconn is a multi-role platform that scans geographic areas, pulls parcel data, analyzes satellite imagery with LLMs to spot improvement opportunities (ADUs, pools, outdoor kitchens), scores each 0-100 and routes it to contractors. Hamza built it end-to-end: a first Next.js version, a v2 rebuild with Stripe subscriptions, direct-mail postcards and AWS Amplify deployment, and, with one collaborator, a Flutter contractor app.

Private client code

Highlights

  • Automated pipeline: scan zones, fetch parcels filtered by value and land-use codes, pull Mapbox satellite imagery, run an LLM assessment, score the lead 0-100, and issue a QR code and shareable link per property.
  • Role-based routing for super admin, admin, support, contractor and homeowner, enforced in Next.js middleware.
  • Admin-managed subscription pricing that creates and syncs Stripe products and prices from the UI, with lead-quota enforcement.
  • Contractor tools: direct-mail lead postcards, support tickets, knowledge base, messaging, activity log, and a processing queue with retries.
  • Started the Flutter contractor app with one collaborator: auth and secure session storage, the contractor and visualization API services, and the Profile and Schedule screens.
  • Prototyped an AI voice sales agent on the OpenAI Realtime API (WebSocket, short-lived session tokens, function tools that hand off to specialist agents), plus a state-wide property export to CSV. Hamza wrote all 29 commits of v1 and all 60 of the v2 rebuild (Jul–Oct 2025).

Technology

Platforms

WebiOSAndroid
Next.js 15React 19TypeScriptMongoDB / MongooseTailwind CSSJWT (jose)OpenAI GPT-4oAnthropic ClaudeGoogle Gemini / Vertex AIRegrid APIMapbox GLGoogle MapsAWS S3AWS AmplifyStripeResendPuppeteerFlutter / DartProvidergo_routerOpenAI Realtime APIPDFKitChart.js

Project visuals

GPT-4o improvement suggestions for one property. Each card gives a placement, footprint estimates from scale cues in the satellite image, a feasibility level and a cost range.

Four improvement cards (Pool Installation, Deck Addition, Outdoor Kitchen, ADU Construction), each with a feasibility badge, location, measurement reasoning, estimated cost range and a Visualize button
Next projectMENA Trade→