Portfolio
MarinadeSolana staking · Product Designer

Redesigned Marinade as the product grew beyond staking

Mobile-first staking, a broader app, institutional decks and interactive product prototypes across two engagements.

$2.5BPeak TVL
6M+SOL staked
2Redesigns shipped
A 2026 position page design for Marinade Native.

A staking app that expected you to know DeFi

The original app showed almost everything it could show. There were a lot of settings, and first-time visitors had to work out where to begin.

More people were arriving without much DeFi experience. They needed a simpler way into a product whose main decision was still quite small: liquid or native staking.

I started as the solo designer on Marinade through an agency, working directly with the founder. I later continued independently and as part of the team.

Start with the phone

The team's analytics showed that more than half of Marinade's users were on mobile. We designed the new app mobile first.

Staking was the starting point. The choice between native and liquid had its own mobile sheet, with the difference between the two visible before committing. I then worked through the supporting pages until the app had moved into the new design.

Earlier redesign: choosing native or liquid staking on mobile.
Earlier redesign: comparing validators across desktop and mobile.

More room for a broader product

After the first redesign, a new Head of Product joined with a broader direction for the app. Marinade needed room for more features and more information.

We began moving to shadcn/ui and widened the layout. Staking and migration sat alongside position management, Recipes and reporting. The interface had to accommodate more of what people could do with their stake.

Alongside the app, I collaborated with another designer on the landing page for what the team called Marinade 2.0.

Later redesign: staking, migration, position management and Recipes on mobile.
A transaction microinteraction concept for Recipes, prototyped in Play.

Institutional decks and StakePoint

During 2024 and 2025, Marinade was doing more work with institutional partners. I worked closely with Hadley Stern on the pitch decks used in those conversations.

My scope also included Twitter assets, merchandise and the visual identity for StakePoint, Marinade's staking event. The same designer working through the app's details was also preparing the materials the team took into meetings and events.

“Harry was instrumental in helping us level up Marinade’s institutional presence. He took something complex — staking infrastructure and DeFi — and turned it into a brand and experience that made sense to enterprise partners. His work gave us the polish and clarity we needed to build credibility with firms like BitGo, Anchorage, and Zodia. He’s fast, thoughtful, and just gets how to make design work for business.”
Hadley Stern

Hadley Stern

Chief Commercial Officer, Marinade

Making the product easier to inspect

When Marinade needed a designer again, the CEO contacted me before starting a new hire. I returned on a contract basis, focused on the product and landing page.

Explore and its detail pages are part of that work: browsing validators, comparing their data and looking more closely at an individual validator. The screens below are final designs; the interactive examples show the behavior work alongside them.

Explore, 2026. Final design from the approved Figma page.
Validator detail, 2026. Final design from the approved Figma page.

Validator pages also got two new widgets. Activity puts a validator's incidents on one timeline: downtime, missed blocks, sandwich attacks and governance votes, each with what it cost stakers. Stakers shows who delegates to the validator, as a table or a treemap. Both are working prototypes, so the behavior was settled before it reached engineering.

Interactive design prototype with sample data. Switch the period, scrub the timeline and open an incident to see what it cost stakers.Open prototype ↗

One position page for every product

Native staking, liquid staking, a single validator and the USDC vault each had their own position page, built at different times. I brought them onto one template: the same position card with total, rewards and net APY, the same chart and period switch, the same action panel and the same transaction list.

Each product changes only what is actually different: the token, the action names and the badges that apply. This work is designed and not shipped yet.

Native, Liquid, a single validator and the USDC vault on one position template, 2026. Designed, not shipped yet.

Work through the behavior, too

My newer work also includes the tooltip system, badges and slider. These are shared parts that appear throughout the app.

The tooltip prototype lets me check what a static frame cannot show: moving between neighboring values, keeping an explanation open while reading it, and inspecting chart points with a mouse, keyboard or touch. Turn on Timing to see the panel I tuned it with: the first-hover delay, the warm window that lets the next tooltip open instantly, and the exit fade.

Interactive design prototype with sample data. Sweep across neighboring labels, rest on a rich tooltip, scrub a chart. Turn on Timing to watch each delay and transition as it happens.Open prototype ↗

Badges went through the same process. The verified and protected-rewards marks have to fit in cards, tables and beside APY values without crowding a validator's name. The prototype compares the three placements from Figma, and each badge expands on hover to say what it means.

Interactive design prototype with sample data. Switch between the three placements, then hover a badge to see it expand.Open prototype ↗

Rules for numbers and layout

Some of the most useful work has no screen of its own. I wrote down two sets of rules the app was missing, both as proposals for the team to review.

Numbers. SOL shows four decimals and USD two, because one step of 0.0001 SOL is about a cent. Tiny amounts never show as zero. Balances round down and debts round up, so the app never shows more than someone can take out or less than they owe. Each rule sits next to the screens it would fix.

Layout. I took an inventory of the spacing on every app page, listed where pages disagreed, and turned it into one set of defaults: content widths, margins, three gaps (12, 24 and 40), and card radius and padding for desktop, tablet and mobile.

Number rules, 2026. A proposal for review, with each rule checked against the live app.
Layout defaults, 2026, drawn from a spacing inventory of every app page.

Shipped twice. Invited back.

  • Both app redesigns from the first engagement reached production.
  • The work covered the product, landing page, institutional decks and event identity.
  • Marinade's CEO contacted me before starting a new designer search in 2026.

Very little of the early interface remains unchanged. Marinade has moved on, and the design has changed with it. I now work with Albert on the product and landing page, with much of the newer work still on the roadmap.

$2.5BPeak TVL
6M+SOL staked
2Redesigns shipped
“Working with Harry has always been an absolute pleasure. His UX skills have significantly improved the user experience of our products. Harry’s collaborative approach and professionalism have made him an invaluable member of our team. I highly recommend him to any organization seeking top-tier UX/UI design expertise.”
Robert Tarabcak

Robert Tarabcak

Head of Product, Marinade

Tell me what you're building

Send your product and I'll record a Loom with the highest-leverage fixes in 48 hours. Free, and no pitch.
Or if you only have the idea, I'll gladly help you with the design strategy on a quick session.

  • Video review in 48h
  • No charge, no strings attached
  • 2 slots open for November
Wanna talk first? Say hi on Telegram

Tell me about it

Takes 30 seconds. I read every one myself.