Products
I've Shaped
I wasn't the founder, designer, or developer. My role: find friction, understand users, and surface what was slowing activation, trust, and retention — then hand off clear recommendations.
Reviewed onboarding internally at Refine Studio — QA, feature recommendations, and simplifying the first-time user experience for a DeFi lending protocol.
- Too many steps before users could do anything useful
- Value wasn't surfaced early enough to hold new users' attention
- Key actions were buried with no clear path to them
- Add a progress indicator — users need to know how close they are to the end
- Surface the top APY on the first screen, not buried in the product
- Cut the steps to make a first deposit to under three clicks
Reviewed the multi-asset swap interface at Refine Studio — focused on the first-use experience and where new users were getting lost.
- No obvious starting point — interface assumed familiarity users didn't have
- Users had to figure out the UX pattern before they could see any value
- A guided first-swap flow would have done a lot of the heavy lifting
- A "Start your first swap" wizard — guided, not free-form, for first-timers
- Show expected output before users confirm, not after
- Surface 3–5 popular trading pairs upfront so users aren't starting from zero
Reviewed trust signals for an AI agent that autonomously trades on your behalf — a product where trust is the entire conversion problem.
- Users had zero visibility into what the AI was doing or why
- No way to audit past decisions — confidence-building was impossible
- Trust is the conversion driver for anything touching funds
- Add a decision log: every trade should have a one-line "why" explanation
- Let users set a risk cap and visibly reflect it in the AI's behavior
- Show live position rationale in plain language, not just raw P&L
Benchmarked their perpetual DEX against dYdX and Hyperliquid — identifying where the experience gap was largest and which fixes would close it fastest.
- Experience gaps vs. dYdX and Hyperliquid were easy to spot — users would spot them too
- Mobile friction points top competitors had already solved
- Order flow was longer than it needed to be
- Reduce the open-position flow to two taps on mobile, matching Hyperliquid
- Add a quick-close button directly on the position card
- Prioritize the highest-gap UX improvements based on competitor parity analysis
Worked inside the platform on growth and onboarding strategy — focused on activation, retention, and making the earn loop clear enough to drive repeat participation.
- Core value loop wasn't clear on first visit — users didn't know what they'd earn
- Active campaigns were buried, first participation required too much effort
- Nothing pulling users back after their first campaign
- A guided "your first campaign" flow — first participation in under 2 minutes
- Surface active campaigns on the home screen, not three clicks deep
- Add a streak or comeback mechanic to create a reason to return next day