My own music, written and produced by me

Mehdi Maiez

Senior Product DesignerAI-Driven Product Experiences

I design complex digital products from problem to working prototype — using AI and code to move faster from idea to validation.

See how I work

Mehdi Maiez — Senior Product Designer — AI-Driven Product Experiences

Scroll to explore

I don't just design screens.I solve product problems.

What I actually do

From discovery and UX strategy through to prototyping and production-ready experiences, I work across the whole product process — using AI and code to explore more directions, validate them faster, and close the distance between a design and the thing that ships.

What I've owned —and what shippingit actually took.

Selected work

See moreshipped work

See all work
PhilosophyDesign → Ship → MeasurePractising since 2012

Proof of Work

I design the whole path — research, system, interface, and the handoff that survives contact with engineering. Then I build enough of it myself to know the design holds. Almost fifteen years of that, most of them leading the team doing it.

Read the long version

A portfolio is not a gallery. Every screen here shipped, met a real constraint, and can be opened by a real user.

The receipts
15productsowned end to end, from the problem through handoff
10shipped & liveopen them yourself — every link is on the project
09industriesfrom MedTech and fintech to cybersecurity
14yearsalmost fifteen, most of them leading the team

Five things I amactually hired for.

What I bring

Not a skills list. These are the parts of the job I own when a product is ambiguous, complicated, or both.

  1. 01

    Product thinking

    Turning an ambiguous problem into a clear product direction — and being able to say what the product should not do.

  2. 02

    UX strategy

    Structuring complex workflows around what people actually need to get done, under whatever constraint the domain imposes.

  3. 03

    AI-driven design

    Using AI to accelerate research, exploration and prototyping — more directions examined in the time one used to take.

  4. 04

    Rapid prototyping

    Getting from concept to a realistic, interactive experience quickly, so a decision is tested rather than argued about.

  5. 05

    Design to production

    Working close enough to the code to know the designed experience survives implementation — because that is where most of it is lost.

From idea toworking product.

How I build

I turn design decisions into working prototypes quickly, so the experience gets tested in something close to the real product rather than a clickable mock. The pipeline underneath joins Claude, Figma, GitHub and Vercel, so design and code read from one source — at Neoshore that cut time-to-market by 60–75% and handoff errors by over 85%. The speed comes from removing the handover, not from skipping the parts that make code maintainable.

The process
  1. Design
  2. Prototype
  3. Validate
  4. Ship
What it runs on
Design
Figma — design system, components, prototypes, with a Figma-to-code handoffFigma
Frontend
React · Next.js · TypeScript · TailwindReactNext.jsTypeScriptTailwind CSS
Backend & data
Supabase or Firebase — Postgres, auth, storageSupabaseFirebasePostgreSQL
Deployment
VercelVercel
AI layer
When the product needs it: LLM integrations (Claude, GPT, Gemini), RAG on private data, voice (ElevenLabs), automations (Make / n8n)ClaudeGoogle GeminiElevenLabsMaken8n

Not vibe-coded.Product-driven.

The difference

I don't use AI-generated code as a shortcut around design. I use it to bring an idea into working form faster, test the real interactions, and find the problems while they are still cheap to change. What comes out the other side is complete — the product, the system it runs on, and enough written down that someone else can carry it.

01

A product that runs

Deployed to production, opened by real users — not a local demo.

  • Authentication and base roles
  • Core features scoped at kickoff
  • Error handling and responsive UI
  • Production stack: React · Next.js · TypeScript · Tailwind · Supabase/Firebase · Vercel
02

A system the team can extend

A visual direction built for the product and its market — not a default style.

  • Precise tokens: color, type, spacing, shadow, radius
  • Dark mode
  • Components delivered in Figma and integrated in code
  • Consistent and extensible — not default shadcn
03

Decisions someone else can act on

So the team — or the next designer — picks up exactly where I left off.

  • Architecture diagram
  • User flows
  • Specs ready for Notion / Confluence
  • Handover guide
Approach
“Start from the constraint, not the moodboard. The interesting design decisions are the ones a real limit forces you to make.”
01

How I start

Research & framing

01 — 03
Contact

Have a complex product problem?

I'm interested in products where UX, technology and new interaction models intersect.

Start a conversation