Role
Team
Timeline
Platform
[The Problem]
Traditional recipe apps strip away cultural context, leaving home cooks feeling disconnected from the global heritage of their food.
[The Solution]
A map-centric mobile experience with two core flows — Cook and Learn — so users can cook a recipe and understand the world it came from.
[My Impact]
Led the full design cycle. Owned 95%+ of the UI. Pioneered a custom AI-driven workflow for authentic imagery and bespoke design tools.
[The Problem]
Home cooks are outgrowing recipe apps.
Three food-lovers asked themselves a question at the start of 2026: why does it feel like there's a gap in the food-education space? People cook at home regularly, often inspired by global travel. Yet the apps they reach for — timers, ingredient lists, step-by-step guides — tell them nothing about why a dish matters, or what it means to the people who've cooked it for generations.
We asked ourselves: What if we created an app that made people feel like they were travelling the world through their kitchen?
How might we foster a deeper connection between home cooks and global cultures by centring recipes around their historical and regional roots?
[Research]
We spoke to real food lovers
I led the interview guide, refined it after INT001, and ran all subsequent interviews. I also designed and distributed the broader survey to 30+ participants — we wanted an organic picture of how people feel when they cook and what they cook.
Affinity mapping across all data narrowed us down to three key insights that directly informed our design decisions.
said cultural context is important to their enjoyment of food
cook at home 3–5 times a week
participants surveyed across interviews and online
Competitive analysis of 20,000+ cooking apps on the iOS App Store confirmed our hunch: no existing platform combines cultural storytelling and regional discovery in an immersive cooking experience.
[Design Process]
Double Diamond, start to finish.
Over 13 weeks I drove the team through both halves of the Double Diamond, moving between collaborative workshops and individual design ownership.
Used AI to synthesise research notes into a persona for "Anya" — our constant reference throughout the project.
A critical exercise in stepping outside my own biases and genuinely inhabiting the mindset of a user who isn't me.
Mapped a day in Anya's life to identify the natural moment when COOKA would earn a place in her routine.
Led the group in building a full story map, then ran a MoSCoW sort to isolate must-haves for our MVP.
Wire-framed three critical flows: Discover content, Explore a recipe, and Cook a meal.
First to sketch initial designs, drawing from All Trails and SoundCloud. Moved into Figma early to establish spacing and style.
[AI Integration]
Building the tools I needed.
I deliberately held off on generative AI until the foundational UX work was solid — I needed to build the muscle, not skip it. Once the research and structure were in place, I stepped into a Creative Director role and used AI precisely where it unlocked something I couldn't do otherwise.
Claude Code
Custom map builder
The maps I found online weren't the right colours. I used Claude Code to build a custom world map creator: full colour control, hi-res SVG export, layers named by country.
Nano Banana &
ChatGPT Images 2.0
Authentic imagery
Drawing on my photography background, I prompt-engineered high-fidelity food and cultural portrait imagery — specific reference material and lighting cues to match Anya's world.
[Visual Design]
From mood board to "Orange Mocha."
My initial designs leaned too heavily styled — early testers said the visual treatment was competing with the food itself. I mocked up three visual directions and ran a simple preference test: "Which of these feels like a worldly cooking app?"
'Orange Mocha' won decisively. That became the pivot point that finally gave the team a shared, buildable direction.
[Final Designs]
The two primary flows
There are two primary flows available to the user;
Cook a meal flow: The user browses a collection of recipes, clicks through to explore a recipe and see related content, then enters "Cook Mode" to see minimal step-by-step screens with method and ingredients.
Explore content via Learn tab: Here the user finds articles, images, podcasts, and more, all related to the region selected and the meals being prepared.
[User Testing]
What real users thought
Several rounds of usability testing on the final designs with participants from our target demographic. Here's what they surfaced — and how I responded.
[Timing & difficulty weren't visible enough]
Users wanted cook time and difficulty directly on the recipe card — no drilling in required. Added those fields to the card surface.
[Cultural articles were buried]
Users wanted quicker access to Learn content. Added a floating dropdown menu for immediate access without disrupting the Cook flow.
[Search bar consumed too much real estate]
Users felt it shrank the map viewport — the most compelling part of the app. Reduced it to reclaim that space.
[Ingredients & Method tabs felt disconnected]
A previous iteration placed the search bar alongside these tabs; it didn't fit semantically. Separating them resolved the confusion.
[Design in the Details]
The decisions a sharp eye will notice.
[Personal Reflection]
What 13 weeks in the Double Diamond actually taught me.
The real growth happened in the spaces between design sprints — with a new baby at home in the mix.
[Balance Act]
Designing a complex product while managing significantly reduced personal capacity exposed a critical risk: over-centralised leadership. When my bandwidth dropped, momentum dropped with it. Going forward I'll build in shared ownership and living documentation from day one.
[Breaking the Deadlock]
We spent weeks circling the "look and feel" without landing. I realised that talking about design wasn't moving us — showing it was. Pitching the map interface and the Orange Mocha theme gave the team something tangible to build toward. Sometimes the most important design decision is just making one.
[Learning to Let Go]
I have a habit of getting pixel-locked on individual components. The team produced its best work when given clear, actionable goals rather than waiting for me to finish something. Moving forward: step back from the screen, keep one eye on the bigger picture.
[The Reality of Research]
Shorter, scrappier testing loops early would have saved days of mid-sprint rework. I'll advocate for faster feedback cycles from the very start of the next project — not just at the end.
