Most portfolios make you hunt. This one reads the question, answers it in writing, and shows the evidence behind the answer — including demos you can run yourself. Everything stays browsable if you'd rather just read.
Or try:
Independent products keep the work honest — commercial reality, technical constraints, and users who leave if it's slow. Two of these have demos you can run. They're simplified versions of the real logic, not screenshots.
How do you know a sale is actually a sale?
A deal-discovery product for the UAE built around source verification: price history, retailer confirmation and store-following, so the feed shows offers that hold up rather than whatever shouts loudest. Designed and built end to end, including the affiliate commerce side.
Turn repetitive marketplace operations into rules.
Seller-side pricing automation on Cloudflare and the Amazon SP-API: scheduled price rotation, per-SKU floors and ceilings the rule can't breach, full history, and manual override. The design work was mostly about making an automated system legible enough to trust.
9.84 AED · floor 9.60
What if people stated a goal instead of navigating a menu?
A conversational navigation model for complex digital services: a person says "take me to my full fines history", the system resolves that to a service state and opens it directly. Menus become the fallback rather than the primary route. The pattern came out of years working on complex service journeys, where the structure is often the problem.
You're using it. The console above is the same pattern applied to a portfolio: your intent is read, resolved to a topic, answered, and the rest of the page reconfigures around it.
Runs entirely in the browser. No request leaves the page.
Small experiments that keep me close to the edge.
Where I test ideas before they show up in client work: interaction studies, voice-first utilities, local AI inside my own build and review process, and developer tooling that removes the chores nobody wants to own.
I work across several layers of a product. Rather than rate myself out of ten, here is each layer paired with something you can inspect or try.
Understanding the system and the commercial constraint before anything gets drawn. Most broken front ends are a symptom of the back of house.
Sale Scan's verification logicJourneys and blueprints for services with real institutional complexity behind them — across large-scale digital products and services.
Intent-based servicesIntent, state and action treated as design material, instead of a chat window parked beside a screen that already didn't work.
The console on this pagePrototypes with real data, real API calls and real failure states — close enough to production that engineering can argue with them properly.
The price rotatorFlutter, Ionic, Angular. Enough to build and ship an interface myself, and to know exactly where that should stop.
Product labREST, Supabase, Cloudflare Workers, Amazon SP-API. The repetitive middle of a process, turned into rules with guardrails.
QualityNinjaVisual and interactive design. Craft first: type, motion, and the habit of caring how things are made.
Interaction systems across web and mobile, and interface behaviour that got steadily more complex.
End-to-end product UX — marketplace work, hybrid apps, Angular and Ionic implementation, API integration.
Product leadership across complex, multi-service digital platforms: design-to-delivery ownership, accessibility, prototyping, and AI-native interaction.
Ambitious digital products, AI-native interaction, automation, complicated service journeys, or a role where design judgment and technical understanding have to sit in the same person. Any of those.