ZAR · Head of Product Design · Nov 2024 – Present
110 countries. 300K users. Backed by a16z.
Brandon Timinsky — SadaPay's CEO — called me back for a second run. The product was already bleeding: five freelancers deep, no design system, no voice, no direction. Twelve months later ZAR is a $20M-funded stablecoin wallet, live across 110 countries, with a design team that runs without me in the room.
ZAR is a stablecoin wallet — digital dollars on Solana — built for the markets where inflation, capital controls, and bank closures make holding local money a losing bet. Nigeria. Argentina. Turkey. Pakistan. 110+ countries in total.
The problem isn't crypto. The problem is trust. Previous crypto products in these markets scammed people; the category starts at zero. Every visual decision — fonts, weights, grid, motion — had to earn back belief that a stablecoin wallet would not disappear overnight.
Rock. Solid. Dollars. Nothing flimsy, nothing bouncy, nothing that could float away.
From a freelancer's leftovers to a $20M wallet, in five chapters.
Brandon Timinsky — my CEO at SadaPay — called me back. ZAR had been running on freelancers: five of them, each pointing the product in a different direction. No design system. No voice. A brand that hadn't settled. Figma files that read like five startups stacked on top of each other.
I took the job as Head of Product Design with one condition: rebuild from the ground up. The first month was not a screen — it was an inventory of what had to be thrown out.
V1 landed at around 200 components and failed within a quarter — product scope shifted, component patterns broke, coverage collapsed. V2 fixed the structure but the brand was still moving underneath it. V3 was the one that stuck.
670+ components. 37 Figma pages. 103 icons. Three platforms — iOS, Android, Web. Token architecture wired to production. Light-mode for browsing, dark-mode locked for amount-entry to build certainty. A tectonic grid where nothing floats.
Techtonic is why ZAR can ship to 110 countries without 110 engineering teams.
We partnered with Ragged Edge in London — the agency that rebranded Wise. I was the client-side lead, translating every brand decision into product reality: the typefaces engineers would implement, the motion curves the animations had to hit, the grid the components would sit on.
EK Modena Heavy for hero numerals — the weight was deliberate. Baikal for body, grounded and quiet. Gold reserved for primary CTA only; nothing decorative. The treatment matches the thesis — rock, solid, dollars — and it carries from marketing through to the moment you confirm a payment.
The system that scaled ZAR across 110 countries.
A tokenized, atomic system — colour, type, spacing, radius, motion — all resolved upstream in Figma and wired downstream into Swift, Kotlin, and React. Every component encodes a constraint of the thesis: gold for primary only, dark for amount-entry, tight margins, no floating.
Three rebuilds are not three failures. They're what it takes to compress a brand, a product, and a regulated flow into one substrate. Once Techtonic settled, feature velocity doubled.
His progression over the years was remarkable.
Brandon Timinsky · CEO, ZAR & ex-CEO, SadaPay
Onboarding was the one I watched most closely. On-ground research in the target markets showed where people quit — ID steps that asked for the wrong document, friction that read as a scam, a flow designed for a western user who didn't exist here. I mapped every touchpoint and stripped it to four. 77.8% fewer steps, measurably higher completion.
Cash QR Consolidation. 72% of transactions were bypassing the order system — merchants were handing out personal QR codes instead of business ones. I consolidated the flow so every transaction routed through the ledger. Reconciled, visible, tracked. That fix protected the business and the user at the same time.
Merchant onboarding went from 19 steps to 4. Same philosophy. Remove everything that doesn't directly serve the user's goal.
I started with Chothu — an internal AI design consultant. Engineers could ask it design-system questions, get component suggestions, and ship standard patterns without a designer in the loop.
Then MACHINA — a five-tool AI suite that automated the repetitive parts of design work. Content generation. Layout suggestions. Component mapping. QA checks. Documentation. Engineer Self-Serve Design stopped being a slide and became a real thing: engineers could ship UI for standard flows without pulling a designer in.
The point wasn't to replace the design team. The point was to take everything that didn't require a designer's judgment off the team's plate, so the team could spend its hours on the decisions that matter. The best design leaders make themselves unnecessary. I meant it.
On-ground research
In these markets, Silicon Valley instincts fail. Two places where service design beat assumption.
Merchants were using personal QR codes instead of business ones — faster to set up, no compliance friction, zero reconciliation. Problem: the business couldn't see 72% of its own volume, and users had no recourse when a payment went wrong.
I mapped the full merchant journey, front-stage and back-stage, and consolidated QR issuance inside the onboarding flow. Every transaction routed through the ledger. Every dispute recoverable. Business visibility and user trust in the same redesign.
The original flow had 18 steps, most imported from a western neobank template. In markets like Nigeria, Argentina, Turkey — users quit by step six. On-ground interviews surfaced the real shape of the problem: wrong documents requested, wrong mental model for "verification," wrong default currency at sign-up.
I redesigned for how people actually behave when they first open a financial app in an inflationary market. Four steps. Same regulatory coverage. Dramatically higher completion.
He understands that the best design work is often the work that unlocks other people. Sebastian Scholl · ZAR
Products are the visible work. This is the quieter one — the operating system that ran the team without me in the room.
Built three times, settled on the fourth attempt. 670+ components, 37 Figma pages, 103 icons, three platforms. Token architecture wired upstream in Figma and downstream into Swift, Kotlin, and React. The substrate that made 110-country coverage possible without 110 engineering teams.
An internal AI tool engineers could query directly — design-system questions, component suggestions, standard patterns — without pulling a designer in. The first step toward self-serve.
Five internal tools that automated content generation, layout suggestions, component mapping, QA checks, and documentation. Took the repetitive work off the team's plate.
Standard patterns engineers could ship without a designer in the loop — gated by the design system, not by a reviewer's calendar. Freed the team for judgment calls.
Client-side lead on the brand overhaul with Ragged Edge (London, rebranded Wise). Typefaces, motion, grid — translated from brand into product in real time, not over a handoff.
Field interviews in target markets — Nigeria, Pakistan, Turkey, Argentina. Journey mapping. Stakeholder research. Every major flow decision had a grounded sample behind it, not a dashboard.
Scaled the design function from me to a capable team. JDs, scorecards, a design challenge calibrated to ZAR's actual constraints, brand-compliance governance that ran after I stepped back.
No handoffs. Design and engineering in the same file, same standup, same review. Token parity enforced in code, not in docs.
Compliance, legal, ops — inside the squad, not on the other side of it. Regulatory constraints entered at research time, not at release time.
Deep dives on the systems inside the story above.
The design system built three times. 670+ components, 37 Figma pages, 103 icons, three platforms. Token architecture wired to production. Once it settled, feature velocity doubled.
Read the Techtonic case study → AI · Deep diveFive AI tools and an internal design consultant. How the design team stopped scaling with headcount and started scaling with automation.
Read the MACHINA case study →How on-ground research in four emerging markets collapsed an 18-step onboarding flow into 4 — without losing a regulatory requirement — and lifted completion across every cohort.
Read the Golden Flow case study →The ledger-consolidation redesign, the merchant flow compressed 19→4, and the AI design consultant engineers talked to instead of the design team. Separate case studies — on the way.
His progression over the years was remarkable.
He understands that the best design work is often the work that unlocks other people. He is practical, technically capable, and genuinely invested in making the product better.