Laravel + PostGIS Property Data Site with Calculators & Lead Routing

I'm building a UK property information website for homeowners and buyers. I have a complete written specification ready, around 55 pages covering database schema, data sources, page structure, design system and performance requirements. The chosen developer gets the full document. Greenfield build, no legacy code. WHAT IT DOES The site combines UK open property data with editorial content and cost tools. Postcode-level sold price data, street sale history and planning application records, all ingested and normalised from free government and open sources. Location-aware cost calculators for home improvements and energy installations. Editorial content in a CMS with location-specific data blocks embedded in articles. User accounts with saved areas and email alerts when new data appears. Enquiry forms routed to third-party partners, plus a paid directory for local businesses. A separate paid toolset for estate agents including image processing and PDF generation. Admin panel for managing all data, content, settings and partners. There are two AI features, both using the Gemini API through a swappable driver interface. The first runs at ingest and rewrites impenetrable council planning descriptions into one plain-English line, stored once and never regenerated, with hard spend caps and graceful failure so it never blocks the import. The second is a drafting tool for estate agents that generates listing descriptions from structured inputs, with strict prompt constraints for UK advertising compliance. Both are metered per call and logged with token counts and cost. This is API integration work with careful cost control, not model training. STACK, FIXED AND NOT OPEN TO SUBSTITUTION Laravel 12, PostgreSQL with PostGIS, Livewire and Alpine, Tailwind, Statamic, Filament, Meilisearch, Redis with Horizon, Stripe, Cloudflare, Gemini API, fal.ai. WHAT I NEED FROM YOU Strong Laravel including queues, scheduled jobs and service-layer architecture. PostGIS and geospatial querying, which is central here rather than incidental. Real experience ingesting and normalising large datasets, since the main sold-price file is around 30 million rows with monthly deltas and there is fuzzy address matching between two separate sources. Server-rendered frontend skills, because there is no SPA framework and the site must work with JavaScript disabled while meeting a hard performance budget of under 2 seconds LCP on data pages. A clean minimalist frontend built from a design direction I provide, not a purchased theme. Technical SEO at scale, meaning programmatic page generation, structured data, sitemaps and meta templating. And an understanding of why a site like this must never create junk database rows, and how to prevent that. DESIGN AND QUALITY BAR Clean and fast is the whole point. The visual reference is a modern restrained dashboard aesthetic, neutral, card-based, one accent colour. Mobile-first, WCAG 2.2 AA, tested on real devices. SCOPE Single delivery, not phased releases. The spec includes a 20-step dependency order. Roughly half the work sits in the data layer before any user-facing feature becomes meaningful. WHEN YOU APPLY, PLEASE TELL ME One, a project where you used PostGIS or handled geospatial data, and what the hard part was. Two, the largest dataset you have imported into a production Laravel app, and how you handled updates. Three, whether you would build fuzzy address matching between two datasets differently to a normalised hash key, and why.