Lead Product Designer
A product designer who builds: 0→1 consumer products, and AI in every step of the workflow.
I’m Mihael Fugaj. I design the flows where people move money and make decisions with consequences, and I build them: prototypes, the production front-end, and the testing in between. 15 years in web, 10 in apps · BA in Design.
- 0→1 Zevo, designed end to end: payouts, deposits and the contactless unlock
- $2M+ pre-seed raised by Zevo, the P2P EV marketplace I designed end to end
- 22 → 0 contrast failures on this site's home page after the accessibility pass I ran with AI agents
How I work
I prototype and test it myself.
Working prototypes over static comps, and I run the user tests. Watching the moment a flow loses someone beats a second opinion on a mockup.
Claude Code in every step.
Research synthesis, prototypes, the production front-end, QA in a real browser. This site is built and run that way.
Figma for the system, code for the truth.
Tokens and components live in Figma. The interface is judged in the browser, on a phone, with real data, not in a frame.
Five pieces that make the case
Each one mapped to what it proves for a lead product design role.
- Zevo 0→1 · money flows A two-sided EV marketplace, designed end to end: payouts, deposits and pricing that owners and guests can read at a glance, plus the contactless handoff where trust is won or lost. Zevo went on to raise $2M+ pre-seed.
- Agents4Fun AI you can trust with value Fair on-chain allowlists, a wallet trust score, and DeBots: 3,500 agent-made bots. Humans and agents use the same contract, fees and odds, and Chainlink VRF draws the winners, so nobody can steer them.
- DePunks Numbers made legible A pixel-punk club on Ethereum. I decided where the website's copy of the chain may be trusted and where decisions must ask the chain: 770 rows = 770 reveals on chain.
- How I build with AI AI in the workflow Dash App, the dashboard I built to run a team of AI agents, and this site built with it: one PM, a roster of 11 subagents, every report backed by evidence. Contrast failures on the home page went from 22 to 0.
- Haas Tooling Design systems · B2B commerce One component system for a spec-driven B2B storefront, with the IA restructured around the buyer's decision sequence: shorter time-to-purchase and higher average order value.
From brief to shipped
- 01 Open the model Map the decision the user is making and the numbers behind it before touching a screen.
- 02 Prototype the hardest moment The step where trust is won or lost gets a working prototype first, not the happy path.
- 03 Test it myself Real people, real tasks, my own notes. Fix the leak, then widen the scope.
- 04 Ship with the code Tokens and components into production alongside engineering, with AI agents doing the repeatable work.
Let's talk
I’m looking for a lead product design role on a consumer product where the numbers matter: money moving, decisions with consequences, and an AI layer that has to earn trust. Full case studies live under Work; the product-management lens is on Product.