Remote King Australia

Problem

Buying a garage or gate remote online is harder than it sounds. Most buyers only have a worn-out remote with no visible model number — they need to identify a compatible replacement from a photo. The existing platform needed better product discovery, clearer navigation, and reliable media storage to turn those confused visitors into confident buyers.

My Contribution

I joined mid-build and took ownership of frontend features end-to-end — from reading the ticket to shipping to production. Specifically:

• Rebuilt product filtering and pagination components to reduce browsing friction across the catalog
• Integrated AWS S3 (ap-southeast-2 region) for product image storage, replacing a previous approach that wasn't scaling cleanly
• Shipped improvements to navigation flows so users could move between product categories without losing context
• Worked in a PR-based preview workflow where every branch gets a Vercel URL before anything merges — a discipline I want to carry into every team I work in

Architecture

The stack is Next.js App Router with TypeScript end-to-end. Catalog pages use Server Components so search engines get full HTML while client-side filters and cart interactions stay snappy. MongoDB handles the data layer. Media flows through AWS S3 with pre-signed URLs for secure, direct browser uploads. Deployments hit Vercel with preview environments per branch.

The Server Component / Client Component split was the most interesting architectural decision — keeping catalog pages server-rendered matters for SEO in a niche where buyers are Googling specific part numbers, not browsing categories.

Outcomes

Features I built are live and serving real customers. The AWS S3 integration is handling production image traffic. Working on something where a broken deploy costs real users was a different standard of care than any side project — it sharpened how I think about testing before pushing.

Learnings

An existing codebase in a live environment is a different discipline from greenfield. The first two weeks were spent understanding what was already there before touching anything. That patience paid off — I was able to take ownership of features without breaking what was already working.