explorer

Evidence over promises.

Every project below is real work, built, shipped, and running. No mockups, no concept art, no "coming soon."

01

Tradelynq

Marketplace Platform
Next.js 16TypeScriptSupabaseStripeResendVercel
Problem

Trinidad's professional services market runs on word-of-mouth and WhatsApp groups. No structured way to find, compare, or trust local professionals across beauty, trades, creative, health, or tech.

Constraints

Caribbean payment infrastructure is fragmented — Stripe covers some, but WiPay and FAC are needed for local adoption. Professionals resist onboarding unless the free tier is genuinely useful. Trust signals matter more than features in a market where everyone knows everyone.

Approach

Two-sided marketplace: professionals pay for visibility and tools, customers browse free. Supabase RLS for multi-tenant data isolation. Verification flow with document upload and manual review. Stripe for international payments, local gateways on the roadmap.

Outcome

Live marketplace with verified professional profiles, category search across 20+ service types, and a payment-gated premium tier. Deployed and processing real enquiries.

View project
02

DevDash

Developer Tools
ElectronTypeScriptReactViteVitest
Problem

Managing multiple AI coding agents, terminals, and project contexts means constant window-switching. The developer's workspace is fragmented across tabs, terminals, and browser windows.

Constraints

Electron apps are notorious for memory bloat. Terminal emulation needs to be native-feel, not a web pretender. The app has to handle multiple concurrent agent sessions without locking up.

Approach

Electron desktop app wrapping a visual command center. Multi-pane terminal with agent session management. Process orchestration across concurrent Claude Code instances.

Outcome

Working desktop app for agent operations — visual terminal with multi-session management. Internal tool, not yet public.

03

This Site

Product Design & Engineering
Next.js 16React 19TypeScriptGSAPThree.jsSupabaseVercel
Problem

Trinovata's credibility needed to match the quality of work it produces. The old static site had premium animations but no substance — craft without proof.

Constraints

No fabricated social proof. All content backed by real work. Premium animations worth keeping — they demonstrate taste. Architecture needed to support SSR, auth, and content expansion without a rewrite.

Approach

Ported from static HTML/GSAP to Next.js App Router with Supabase auth, preserving every animation while building toward server-rendered content sections.

Outcome

Craft, substance, and context in one site. Deployed at trinovata.com.

View project