Enterprise · Real Estate
Search, listing, transaction and analytics infrastructure for a vertical operator. Sub-200ms median search latency across three cities.
Overview
Vertex operates 4,200 active listings across Toronto, Vancouver, and Montreal. They're a vertical operator agents, photographers, in-house mortgage competing for share against the Zillow-class platforms that own the discovery layer. Their previous tooling was a stitched-together stack of three SaaS products that none of them owned.
Platform-grade search and listing infrastructure with sub-second response times across three regions, on a budget that didn't allow for a five-year build. The team needed to keep all data in their own database both for compliance and because their listing relationships were the company's defensible asset.
Approach
01
We built a typo-tolerant, geo-aware search index sitting in front of Postgres. Inverted indexes for keyword search, PostGIS for proximity, BM25 ranking with locale-specific boosts. Sub-200ms median across all three regions on commodity hardware.
02
A real CMS for the agents listing intake, photographer scheduling, document workflow, MLS sync, change history. The CMS surface was as important as the customer-facing search; agents are the product.
03
Offer management, document custody, and a settlement workflow integrated with their mortgage arm. Audit trail, e-sign, and multi-party concurrency baked in from the first day.
Outcome
Vertex now runs entirely on the platform we shipped. Search latency stayed under 200ms during the 2025 spring-market peak (3.2x normal traffic). The CMS is used by 140 in-house agents and 38 photographers daily.
We can finally tell a story about how we run our listings, instead of apologising for the tools we use.
Stack
Thirty minutes. We'll listen, and we'll tell you plainly whether this is a fit. If it isn't, we'll point you at someone it is.
No discovery deck. No sales call. One conversation.
