Open to Jobs and Collaboration

UX designer.Product strategist

Design that solves,
not just shows.

Four projects. Four domains. One conviction: the best UX doesn't ask users to adapt, it adapts to them.

Telomeres
Intelligent Environmental-Geographical Tracker
Environmental IntelligenceUX Designer5 months · 12-person team
The Problem

Field operators, analysts, and commanders shared one fragmented system - built around data processes, not human decisions. Critical environmental intelligence existed, but no one could act on it in time. The interface was the bottleneck in missions where seconds matter.

UX Insight & Research Direction
Outsider as Advantage

Entering without domain expertise forced the shortest cognitive path. Mapping every term, gesture and decision from scratch revealed where the system assumed knowledge users didn't have.

Three Personas, Three Mental Models

A commander needs situational awareness. An analyst needs data depth. A field operator needs speed. One interface failing all three was the core structural problem.

Data Without Context Is Noise

Users couldn't trust a number they couldn't trace. Research revealed that data pedigree - knowing where a value came from - was the missing trust layer across all roles.

Complexity Staged, Not Hidden

Inspired by game UI: progressive reveal, not progressive hiding. The direction was to stage information by decision priority, not by data hierarchy.

Design Decisions
Role-Driven Flows
Three distinct interfaces for three mindsets - operator, analyst, commander - maintaining one visual system underneath.
Game-Inspired Staging
Borrowed Call of Duty's tactical UI logic: preview data state before committing attention. Tames density without removing depth.
Data Pedigree System
Clear traceability: raw sensor → collected → calculated. Every number has a verifiable origin, critical for high-stakes decisions.
Simulation Mode
Shifted the platform from reactive to strategic - passive forecasting became interactive scenario planning.
Key Insight

Being unfamiliar with the domain was the advantage. Outsider eyes find the shortest cognitive path - stripping jargon to expose what users need to decide, not just to see.

Learning

Design specs need narrative context. Daily standups weren't enough - a dedicated design sync would have preserved intent and reduced friction during development handoffs.

FactWeavers
AI-Powered Self-Service Data Analytics - Client: DBIZ
Enterprise AnalyticsUX Designer & Strategist6 months
The Problem

A powerful analytics platform built by engineers - for engineers. Business users needed 40+ minutes just to generate a single report. It was a Ferrari no one could start. The platform's capability was locked behind its own complexity, and non-technical decision-makers paid the price every day.

UX Insight & Research Direction
User as Research Method

Rather than interviewing users about pain points, I became one. Mapping the platform as a confused end user produced more actionable insight than any survey - surfacing friction that stakeholders had normalized.

Internal Language vs. Learnable UI

The interface used semantic icons, color states, and jargon that made perfect sense to the engineering team and zero sense to business users. Every badge, every color had an undocumented meaning.

Cognitive Load at Entry

A blank workspace with no guidance caused analysis paralysis. Users needed a clear starting scaffold - not a blank canvas - to feel confident beginning work.

Familiarity Beats Cleverness

Users didn't want a new paradigm. They wanted familiar spreadsheet-like flows, enhanced. The design direction prioritized progressive enhancement over reinvention.

Design Decisions
Become the User First
Mapped the platform as a confused real user before designing. The 40-minute struggle became the north star - a felt experience to design against.
Visual Language Redesign
Decoded every semantic icon, badge, and color state - standardized into a coherent design system. Internal shorthand became learnable UI.
Progressive Disclosure
Workspace opens as a guided template. Contextual tools appear as users engage - zero cognitive load at entry, full expert power beneath.
Pre-Launch Rebranding
New logo, color system, typeface - proposed when the product's maturity outgrew its identity ahead of the B2B market launch.
Key Insight

Non-technical background wasn't a limitation - it was the research method. Experiencing the platform as a confused real user surfaced friction that stakeholder interviews alone would have missed.

Learning

Constraints accelerate focus. Fixed architecture forced prioritization of highest-impact UX changes. Launch pressure - when channelled correctly - isn't always the enemy of good design.

ONETemasek
Unified Digital Workplace - Temasek Holdings
Corp Tech / EnterpriseLead UX Designer1.5 years · 34 countries
The Problem

Temasek's workforce juggled 20+ disjointed internal tools, each with its own interface and workflow. Every context switch cost 5-10 minutes. Across 1,000+ employees in 34 countries, that became a silent, compounding productivity drain - invisible to leadership, debilitating to staff.

UX Insight & Research Direction
Perceived Needs ≠ Actual Needs

'Having the latest market index on the homepage would help,' said users. Behavioral data disagreed - it rarely influenced daily actions. Observation, not feature requests, guided every prioritization decision.

78% Search, Not Browse

Usage data revealed 78% of users located apps via search - because the sidebar was frequently cropped or hidden on Windows laptops and mobile. Layout strategy had to follow actual behavior, not assumed behavior.

Familiarity as Feature

Enterprise users expect known interaction patterns. Forcing reinvention would break trust before the platform had a chance to earn it. The design direction was enhancement layered on familiarity, not replacement.

AI Trust Is Designed, Not Given

Investment staff were skeptical of AI outputs. Research uncovered specific trust barriers: hallucinations, lack of source attribution, unclear scope. Each had to be explicitly designed against - not assumed away.

Design Decisions
Prioritize 12, Defer the Rest
Focused on 12 essential micro-apps with highest daily value. Lower-priority tools moved to passive states - reducing noise, preserving attention.
Preserve Mental Models
Preserved core interaction patterns from legacy tools and layered personalization, responsiveness, and AI augmentation on top - not beneath.
Behavior-Led Layout
Reshaped layout around search-first discovery. Minimized visual weight, surfacing only the most relevant content above the fold based on actual usage data.
GenGini - AI Designed for Trust
Integrated secure AI with explicit expectation-setting, source traceability, and limitation transparency. Iterated from 50 to 500 beta users using behavioral analytics.
Key Insight

The biggest design decision was what NOT to build. Users said they wanted market indexes; behavioral data said otherwise. Perceived needs and actual needs are rarely the same - and conflating them is expensive.

Learning

Great enterprise UX isn't about what you build - it's about what you choose not to build, and how you help people navigate change. Underestimating change management cost more than any design misstep.

PAXU
Anti-Trafficking Safety App - Pacific Links Foundation
Social ImpactUX Designer2M+ reached
The Problem

Young migrant workers in Vietnam - with limited digital and economic literacy - were falling for false job offers and fake agencies. They had no accessible, localized tool to verify what they were being promised. Knowledge was the missing layer of protection between vulnerability and exploitation.

UX Insight & Research Direction
Design for the Worst Case

Our users might be reading this app in a high-risk situation, on a slow phone, with little trust in technology. Every interaction had to be legible under stress - no jargon, no confusion, no dead ends.

Low-End Device Reality

Usability testing on actual low-end Android devices revealed that common patterns failed: dropdowns were hard to tap, modals confused users, multi-step flows caused abandonment. One task per screen became non-negotiable.

Abstraction Kills Understanding

Concepts like 'broker fees' or 'salary comparison' were meaningless to users without economic literacy. Research pointed to culturally anchored metaphors - food prices, local goods - as the only reliable entry point.

Community as Safety Net

Users trusted peers more than institutions. Community-validated information - real employer reviews - was more persuasive than official warnings. The platform had to be a community, not just a tool.

Design Decisions
The Phở Index
Wage comparisons translated into bowls of phở - a culturally resonant metric that bypasses economic literacy barriers entirely.
Low-Literacy First Design
One task per screen. Conversational language. Video instructions. Built for low-end Android with unreliable connectivity.
Community Intelligence
Real employer reviews and localized trafficking alerts turned passive users into active protectors within their communities.
Zero-Friction Emergency
National hotlines, police, and NGO contacts always one tap away - never buried in navigation when seconds count.
Key Insight

Design for dignity, not just usability. Every interaction was evaluated through one lens: does this feel safe to a frightened, first-time smartphone user with limited trust in technology?

Learning

This project changed what 'good design' means. It's not how it looks or how it works - it's how many people it keeps safe. Design is a tool for dignity, and that responsibility is permanent.

More below

Design Philosophy

Complexity is not an excuse
for bad UX.

I work at the intersection of dense systems and human clarity - translating geospatial data, enterprise analytics, corporate intranets, and high-stakes safety tools into interfaces people actually trust.

01
Radical Clarity
Simplicity isn't removing things - it's knowing which things matter. Entering a domain as a stranger is the methodology, not the handicap.
02
Constraints as Briefs
Fixed architectures, tight timelines, and legacy systems are not obstacles - they're the frame that forces more focused, implementable decisions.
03
Empathy as Method
Research is a posture, not a phase. Every project starts by becoming the user - experiencing their confusion, their cognitive load, and their risk firsthand.

Capabilities

What I bring

UX Research
User InterviewsStakeholder WorkshopsJourney MappingUsability TestingCo-creationAI/LLM UX evaluation
Interaction Design
Information ArchitectureWireframingPrototypingUser FlowsHuman-AI interactionAI agent interactionAI context engineering
Design Systems
Component LibrariesAtomic DesignFigma VariablesAccessibility
Product Strategy
MVP ScopingFeature PrioritizationStakeholder AlignmentRebranding
Domain Experience
GovTech / DefenceEnterprise SaaSSocial ImpactCorp IntranetAI Products
Collaboration
Cross-functional TeamsDev HandoffDesign LeadFacilitation
Visual Design
High-Fidelity UITypography SystemsData VisualisationRebranding
Tools
FigmaFigJamMiroMazeNotionSharePointCursorClaude CodeAWS Bedrock

Get in Touch

Let's design something that matters.

Open to senior UX roles where design carries real stake - products that reshape how people work, decide, or stay safe. Coffee is always on me.

"Great enterprise UX isn't about what you build - it's about what you choose not to build, and how you help people navigate the change."

- Portfolio 2025

Available for Senior UX Roles
Full-time · Contract · Open to relocate