Hand-built

Alohub

Real-estate web platform · tech lead of a team of seven · 2021 to 2023

A website builder for real estate developers: each client got a public site, an admin app with a quoting tool, an inventory viewer and a sales CRM. I was the senior frontend developer and tech lead of the team behind it.

The problem
One team had to build a different website for every property developer.
What I did
I led a team of seven and built one editing engine for every section.
The result
Real estate developers configured and ran their own websites without engineers.
Built with
ReactNext.jsJavaScriptRedis
RoleSenior frontend · Tech lead
Team7 people
Timeframe2021 to 2023
ResultClients ran their own websites without engineers
75%less development time, from the framework I built
7people on the team I led as tech lead
4 appsper client: site, administrator, inventory and CRM
The problem

Every property developer wanted a different website, and one team had to build them all

Each real estate developer needed its own branding, sections and layout. Building and maintaining a separate website for every client would have tied up the team. We needed one configurable product that could handle everything from photo galleries to payment plans.

How I built it

I led the team and built the frontend

I was the senior frontend developer and tech lead of a team of seven. I wrote the frontend and made architecture calls with the backend team. The platform had two halves: one Next.js app that server-renders every client's public site, and Alohub, the admin app where a client builds that site, quotes a unit, and watches inventory, sales and leads.

One editing engine handles every section type.

Alohub is where a property developer configures its own public website, and every kind of section needed a way to be edited. Instead of a bespoke panel per section type, I built one generic engine: about fifteen reusable controls, a declarative map saying which control edits which field, and one resolver that walks any nested path to reach it. Adding a section type became one entry in that map, with no new screen. That engine is the core of the framework the 75% comes from.

Clients could configure their own quoting rules.

The quoting tool lived inside that same administrator, and it is how a buyer gets a real price for a real apartment. Every developer sold on its own terms: what percentage the deposit was, how long a quote stayed valid, how many installments, how the discount moved with the size of the deposit. I built the tool that had to hold all of those rules and price a unit out of live inventory. A wrong number in a published quote is a legal risk.

The preview uses the same code as the site.

The editing panel sits beside a live preview of the client's real public site, running the same renderer that serves visitors, so what you edit is what ships. I built the bridge on the browser's own messaging between the two frames, debounced 800 milliseconds so the preview stays smooth instead of redrawing on every keystroke. If the preview drifts from the real site, a client ends up publishing something they never actually saw.

The impact

A platform that launched sites in a fraction of the time

The platform ran in production: clients configured their whole public website in Alohub, from the hero image to the payment plans, and managed inventory, leads and sales reports. The generic editor is what made it economical, and the quoting tool is what made the numbers on it trustworthy. A client who wanted a new section on their site got it from the same engine every other client was already running.

What it proves

I can find the one abstraction that removes repeated work for a whole team, build it, and lead seven people to deliver on top of it.

Screens