I notice annoying problems in everyday life, then build and test my way out of them — with AI as my digital collaborator to learn, fail and move forward.
Two decisions on the same surface — the fare a traveller picks, then what they add to it. Someone else’s roadmap, real revenue pressure, and eight carriers that each bundle differently. Each started with a call I had to defend, so that’s what leads.
AlmosaferRevenue wanted the upgrade pre-ticked. I refused the dark pattern and argued the harder position: travellers were losing to airline sites because they couldn’t compare, not because they were cheap. So I made eight carriers’ bundles comparable on six attributes — and let people choose up on purpose.
The brief was attachment rate on trip insurance, stay discounts and eSIM. The constraint I held: no extra could cost a completed booking — so placement, wording and timing were tested against completion, not just attach.
Before jumping to solutions, I want the real goal — what success looks like for the user, the business and the people behind the request.
I look for the real constraint. Sometimes it’s technical, sometimes behavioural — and sometimes we’re solving the wrong problem entirely.
Outcomes, not prettier screens. Real change in behaviour, clarity, confidence or revenue — measured, not assumed.
Since 2024 I’ve worked as an independent product designer — taking products from discovery and PRD through prototyping, AI-assisted build and launch. These are the ones I’ve designed and shipped in that time, most of them live.
Interviewers improvise and pitch it at the wrong level. Candidates guess what’s being assessed. Nobody leaves with what they came for.

5:47pm, a full fridge, and nobody can decide. The obvious build was a bigger recipe library — but more options was the problem, not the fix.

What it is: an app where you find verified personal trainers near you, see their price up front, and book and pay inside the app — plus a separate business tool where trainers manage their profile, availability and bookings. Two apps, one codebase, built in React Native on Firebase.
Why it exists: booking a trainer ran through Instagram DMs and bank transfers. No price to compare, no proof anyone was qualified, no protection if it went wrong. The insight: choice was never the bottleneck — people want to trust the one they pick.
LittleRootsIf you're building something where process is the bottleneck and great design is the unlock — I'd love to help. I work fast, kindly, and with AI in the loop. Bring me your messiest problem.