OKX

Problem

The Trade page was dense, rigid and hard to customize, forcing different traders into the same workspace.

What I did

I combined 34 European KOL interviews, 15 VIP and high-frequency trader interviews, and competitor analysis to identify the biggest gaps.

Work

I led a modular workspace traders could customize and save, bringing AI market summaries, footprint charts, order book heatmaps, derivatives data and opportunity discovery into OKX.

The solution & projected outcome

• +15% 30-day retention
• +25% session duration
• +5% volume per user
• +12 NPS

Problem

European derivatives required eligibility checks, account activation and funding before a first trade. Each step created friction and drop-off.

Research

We interviewed 34 KOLs and traders at the 2026 Barcelona–Catalunya Grand Prix. A key failure: users had funds on OKX but reached trading with zero available balance.

My role

I led design from discovery through compliance, activation, funding, first trade and position management. With Product, Compliance, Legal and Engineering, I turned fragmented steps into one progressive flow and resolved the balance issue.

Outcome

• 120K+ activated traders
• $650M peak daily volume
• 6 consecutive weeks of trading-volume all-time highs

Problem

Mobile dominated usage, but interviews with 34 advanced traders and KOLs revealed distinct roles: mobile for managing positions; web for analysis and complex execution.

What I did

I reframed an underinvested web experience—with overloaded navigation and a marketing-led homepage—as a workspace for analysis, decisions and execution.

Work

Designed a personalized dashboard for markets, assets, positions, watchlists and opportunities, using customizable widgets shared with the Trade experience.

Outcome

A daily trading workspace designed to give advanced traders fewer reasons to leave OKX. Currently in development.

Problem

Perpetual contracts drove 76.3% of OKX’s trading volume, but Advanced mode’s margin settings, funding transfers and trading controls kept many retail spot users out.

Business case

I modeled revenue using OKX’s 15.4%–16.7% EEA Simple/Advanced volume ratio, competitor trading volumes and published taker fees.

Personas

Directional Daniel: A spot holder who wants to trade a market view without Advanced mode’s complexity.

Sideline Sofia: A stablecoin holder reluctant to transfer funds and worried about losses while away.

Outcome

• 100K+ daily active traders
• $300M average daily trading volume
• $208M projected annual revenue

Aligned the four-market MVP scope in one quarter, versus six historically. Codified mandatory stop-loss, 3× default leverage and automatic transfers in the Simple Perps design.

Problem

We aimed to launch regulated retail derivatives under ASIC requirements without fragmenting the global UI. A 2× leverage cap, eligibility checks and projected 95% perpetual-contract volume shaped the rollout.

The solution

I led a two-phase rollout, making eligibility onboarding feel like a product unlock and using Australia to validate a globally scalable architecture.

Outcome

In the first 30 days, open interest grew to $23K+ and position holders from 1 to 8. The deliberately small launch validated the regulated experience and became the foundation for a shared Simple Perps surface across four regions.

Problem

USDG offered 4.1% APY, boosted to 5%, but users only discovered the yield inside Earn—not when depositing.

What I did

I brought USDG’s yield into the deposit flow across platforms, targeting 10× AUM and +15% CTR.

Outcome

The integration delivered 15× USDG AUM and +32% CTR, increasing Earn revenue.

Problem

In research, 60–70% of retail users named lack of understanding as their top blocker. Standalone chatbots couldn’t solve a fragmented experience.

Six pillars

Adaptive Personalization, Proactive Guidance, Predictive Analytics, Progressive Automation, Portfolio Intelligence and Behavioral Learning.

Solution

• Data layer: Exchange, wallet data across 70+ blockchains and third-party feeds in one schema.
• Intelligence layer: Six engines built on intent, reasoning, memory and learning.
• Presentation layer: Eight adaptive product surfaces.

Outcome

Established OKX’s shared reference architecture for AI, reducing duplicated work and shifting the product toward proactive financial guidance.

Problem

Over 70% of mobile users opened Exchange Mode daily, but its five tabs had lost clear roles and consistency.

What I did

Using my User Feedback Intelligence Dashboard, I turned 3,100 tickets into 22 priority problems for one coordinated release.

Outcome

• Projected −35% Exchange Mode support tickets within 90 days post-launch.
• “One tab, one job” adopted for Pay, Earn and Invest revamps.

Problem

Studying advanced traders isn’t the same as experiencing their workflows.

What I did

I hosted hands-on trading masterclasses for Design, Research and Product, connecting market structure, price action and liquidity to real trading decisions.

Work

We traded together using volume, open interest, footprint charts, DOM, order book heatmaps, CVD and Fibonacci levels to understand trader positioning.

Outcome

Two sessions—in person and remote—gave teams practical experience with the tools and workflows they design for.