Skip to content

Selected work

Case studies, not screenshots.

Each project is documented the way a business would evaluate it: the problem, the evidence, the scope decisions, and what actually changed. All three are live — go use them.

Fintech · SME tooling

Tally

Financial clarity for small businesses

FlagshipLive2026
My role
Business analysis & requirement elicitation
Product discovery & strategy
Product design (flows, wireframes, UI)
Product engineering
UAT & release validation
Next.jsTypeScriptSupabasePostgreSQLFigma

Business problem

SME founders cannot answer basic questions about their own business finances. Existing bookkeeping tools are accountant-first: dense ledgers, unfamiliar vocabulary, and a setup cost that owners abandon before they see value.

Research & discovery

I interviewed small-business owners across retail and services to map how they actually track cash flow — notebooks, chat threads, memory — and where decisions break down at the end of each week.

Requirements

Hypotheses were tested with prototypes before scope was locked. Findings converged on a lean requirement set optimised for clarity over completeness: plain-language insight, near-zero data entry friction, and one screen that answers 'am I fine this week?'.

Solution

I defined and shipped a six-week MVP: a streamlined dashboard, plain-language insights instead of accounting jargon, and a low-friction entry model that fits an owner's real routine.

Business impact

20 active testing users onboarded with repeat-usage patterns and a strong qualitative demand signal from the target segment — enough validation to justify continued investment.

20

Active testing users

6wk

Discovery to live MVP

1

Screen to answer the core question

Consumer · Discovery

Zemo

A gift discovery experience built around intent

Live2026
My role
Product strategy & positioning
User research & requirement definition
Experience design
Product engineering
Next.jsTypeScriptSupabase

Business problem

Choosing a meaningful gift is a high-effort, low-confidence decision. Marketplaces optimise for catalogue breadth, which leaves the buyer with more options and less certainty.

Research & discovery

Lightweight user research into gifting behaviour surfaced the real job-to-be-done: buyers are not shopping for products, they are looking for confirmation that a choice will land emotionally.

Requirements

I scoped the MVP around three things only — guided discovery, personalisation, and shareability — and explicitly cut catalogue depth, accounts, and social features from v1.

Solution

Launched an MVP centred on a fast, curated discovery flow that narrows options quickly rather than expanding them.

Business impact

Meaningful first-day traction and early adoption from the target audience, validating guided discovery as the product's core mechanic.

3

Requirements kept in v1 scope

Day 1

Adoption from target audience

Civic technology

Ease Voting

Election transparency people can verify

Live2026
My role
Requirements engineering across three personas
Product strategy
Product engineering
Stakeholder management
ReactDjangoPythonPostgreSQL

Business problem

Voters and observers lack a trusted, real-time view into election processes. Without verifiable visibility, confidence in outcomes erodes regardless of how the process was actually run.

Research & discovery

Interviews with voters and civic-tech stakeholders isolated where trust actually breaks: not in the count itself, but in the silence between the count and the announcement.

Requirements

I framed and tested hypotheses around credible transparency, then gathered requirements across three distinct personas — observer, voter, and administrator — each with different trust thresholds.

Solution

Defined the product strategy and led product engineering for a platform that surfaces verifiable, real-time election data to every persona in a form they can independently check.

Business impact

Delivered a working platform positioned as a reusable civic-tech foundation for transparent elections rather than a single-election tool.

3

Personas modelled end-to-end

Real-time

Verifiable data surface