Hii, I’m Pelumi, a product designer with a growing focus on design engineering and AI. I design end-to-end experiences, build the systems behind them, and use AI-assisted workflows to explore ideas, solve problems, and turn concepts into working products.
Case Studies
Kuda · 2026 · Mobile design
Kuda App Redesign
Redesigning the app’s navigation and visual experience to bring Kuda’s new identity into the product.
Bitescout is a mobile app concept for finding a place to eat and keeping the places worth coming back to. I designed the experience from competitor research and user flows through to an interactive prototype and usability testing.
Company
Independent
Role
Product Designer
Type
Mobile Product Design, UX Design, User Research
Impact
Validated the search and save flows with 13 Maze responses
The core experience connects discovery, restaurant details, and saving a place for later.
Too many places to look, no easy way to choose
Restaurant recommendations are scattered across social accounts, review platforms, and maps. Finding somewhere appealing is only part of the task: people also need enough information to decide, and a way to keep places for later.
The assessment asked for a mobile app that helps people discover and save restaurants nearby, with quick meal decisions at its centre. I focused on bringing discovery, evaluation, and saving into one connected experience.
What I took from existing products
I explored TripAdvisor, Resy, Airbnb, Google Maps, and Reisty to understand established discovery patterns. Cuisine, price, and rating filters were common, as were favourites and curated lists. Reisty’s use of trends and topics offered another way to help people find places beyond a direct search.
This was personal research and competitor analysis. It helped define the design direction; it was not a substitute for testing the experience with users.
Make the essentials visible. Show enough information to compare restaurants at a glance, with a clear route into the details.
Support different ways of searching. Let people browse nearby places or narrow the options by cuisine, budget, and dining preferences.
Make discoveries useful later. Help people save restaurants into personal collections and choose who can discover their lists.
Connecting discovery to a saved place
I mapped the steps from finding a restaurant to adding it to a list. The flows accounted for different entry points: nearby browsing, cuisine searches, and community recommendations. The core journey needed to work whether someone arrived with a specific plan or was still looking for inspiration.
The flow from discovering a restaurant to reviewing its details, then saving it to a personal list.
Start exploring without signing up
The concept opens straight into nearby restaurants without requiring signup. Map and compact views support two ways of comparing the same options: where a place is, and what it offers.
In map view, numbered markers indicate relevance, while a distinct selected state connects the map to a carousel of restaurant cards. People can explore through either the map or the carousel.
The compact view puts restaurant information into a scrollable list, with quick filters and a shortcut back to the map. Saving is available directly on a card, so a promising place does not require a detour through its details.
Search starts with the context of the meal, then lets people refine the results with filters and sorting.
Search around the meal someone has in mind
I designed search around phrases such as “Seafood on Friday night,” rather than cuisine keywords alone. The intended interaction captures more of a person’s dining plans and uses that context to frame the results. This is a proposed search behaviour in the concept, rather than a claim about a working search engine.
For people who want more control, advanced filters narrow the options by rating, cuisine, price, and restaurant type. Sorting by relevance, distance, or rating offers a further way to compare the remaining choices.
Restaurant details bring reviews, menus, and practical information into one decision surface.
Give people enough context to decide
The restaurant detail page brings reviews, menus, and practical information into one place. Tabs make the sections easier to reach, while save and share actions remain part of the decision flow. The goal was to let people move from an interesting recommendation to an informed choice without switching between sources.
Turn a recommendation into a collection
Lists give saved places a purpose: a food tour, favourite spots, or somewhere to try next. The save flow offers “My List” as a default destination and allows a restaurant to belong to multiple lists.
People can also create a list while saving, rather than leave the task to organise things first. The create-list screen includes a character limit and a choice to make the list discoverable to the public or to followers.
The save flow supports a default list, multiple lists, and creating a new collection in context.
Testing the core journey
I ran an unmoderated Maze usability test with 13 respondents. The test covered searching for a restaurant, reviewing restaurant details, and saving a restaurant to a list.
4.7/54.7/5
Average ease rating for search (13 responses)
4.5/54.5/5
Average ease rating for saving (13 responses)
11/1311/13
Rated saving 4 or 5
The live report shows strong perceived ease for both core flows: search averaged 4.7 out of 5, while saving averaged 4.5 out of 5 across 13 responses. Eleven respondents rated saving 4 or 5, while two rated it 3. Read the full results in the Maze report. With a small test group, these findings are directional feedback rather than evidence of performance in a launched product. The report does not establish gains in retention or restaurant visits.
What I would improve next
The gap between search and saving ease makes the save-to-list journey a priority for further review. My next step would be to inspect the test’s friction points, refine the interactions, and test the changes.
The concept also leaves room for a Discover experience built around real dining experiences and community recommendations. That could help people assess recommendations with more context, but its effect on credibility and repeat use would need to be tested.
Kuda.
