UniLink: mentorship for exchange students — Daphne Cheng · Daphne Cheng

Talk to a local first. Studying abroad should feel like an adventure — not figuring out how to turn off a fire alarm alone.

Overview

UniLink is a mentorship platform that connects incoming exchange students with peer mentors at their future university — for academics, culture, and the messy first 30 days abroad.

I owned visual design, the design system, interaction design, and prototyping on a three-person team over four weeks.

role: Visual design, design system, interaction design, prototyping
team: Ruhin Gharai (PM), Brian Camilo (research), me (visual)
timeline: 4 weeks
methods: Community survey (60+), 4 interviews, competitive analysis, heuristic walkthroughs with 8 students
tools: Figma, FigJam, Google Forms

UniLink overview: phone collage of app screens with Situation, Task, Action, and Results — Temp archive — original overview layout

The problem

Isolation in a new system. 1M+ exchange students each year land without familiar support for culture, logistics, and daily life.
Scattered help. Information lives across university sites, group chats, and random forums.
People want people. Nearly 4 in 10 said they'd rather just talk to someone.

Here are the biggest complaints from a one-week community survey (60+ responses), then 4 interviews and a competitive scan:

Information is scattered across multiple platforms

Difficulty navigating logistics in a foreign environment

39% of global students would prefer to just talk to someone

Problem section with How Might We prompt and three research insight cards with charts — Temp archive — problem + insight cards

Discovery

Not more resources, tools, or systems. The best way to show the how is talking to a human.

We scoped to exchange students (not all housing searchers), then cut features hard for a four-week build.

Feature	Keep?	Why
Q&A forum	Yes	Familiar (Piazza/Quora); real human answers
Community events	Yes	Online → real-life involvement
Video call	Yes	Safer after matching
AI chatbot	No	Hallucination risk; shifts away from peer help
Financial proof	No	Fintech scope creep
Housing match	No	University's job
Emergency contact	No	Crowded; not our edge

Mentorship over housing search

Research widened the problem past housing. Narrowing to exchange students let us build a mentorship platform that matched what people actually asked for.

why

Housing is a university's job. Mentorship — talking to someone who's been there — was the gap competitors weren't filling well online.

impact

A tighter product: match, book, ask, and meet — instead of a half-built housing marketplace.

gave up our first idea — roommate / housing match

Discovery: insight callout, personas, feature prioritization table, and low-fi wireframes — Temp archive — discovery + wireframes

Design

I mapped information architecture with the team so we knew how many screens we needed, then iterated hi-fis in the open so hierarchy read from across the room.

Open a craft thread when you want the detail — keep the main scroll light:

1/ Building a design system

Don't try to finish the system in one shot. We started partial and filled tokens and components as screens landed — then the library kept the product cohesive.

System at the end — tokens + components

2/ Animating a logo

Early mono wordmarks read like buttons. We moved to a connected wordmark and a 3–5s intro (dot → line → “university link” → icon) so the brand landed before the first screen.

Ideation → Smart Animate intro

3/ Iterating on the navigation bar

The tab bar is on every screen. Color-block active states were clear but loud — we switched to a top indicator line + purple active icon so the page stays primary.

Navigation bar rationale and iterations from color-block to indicator-line active state — From color block → indicator line

Six navigation bar accessibility variants covering color-blindness, visual cues, contrast, and feedback — Accessibility variants

Test with users

I led heuristic walkthroughs with 8 global students — hunting logic fails, then revising.

1/ Confusing Q&A layout

The Q&A page design can be better, and I think instead of filter maybe sorting to pick category questions would be easier

— exchange student 1

Issue: Q&A flow wasn't intuitive — too many steps before finding answers.
Why: Filter-heavy layout asked for work before value.
Fix: Restructured toward a Reddit-like feed: clearer posts, sortable categories, Create Post up front.

Before and after Q&A page redesign on mobile — Before → after Q&A layout

2/ Booking icon is hidden

The booking symbol for scheduling was hard to find and I would rather have it on the navigation bar as I would check it out often

— exchange student 2

Issue: Booking was hard to find.
Why: Top-bar icons read as secondary; outside the thumb's normal path.
Fix: Moved bookings into the bottom nav — where people already check often.

Booking icon moved from top bar to bottom navigation — Top bar → bottom nav

3/ Saving an event

I want to be able to save events I want to go maybe so I can apply closer to the date once I know my schedule better

— exchange student 3

Issue: Users wanted to save events for later.
Why: They expected more actions on cards — not just view and go.
Fix: Reused the mentor-card star component on event cards for a quick save.

Event card with star save control called out — Star save on event cards

Refine

I led mid-fi → hi-fi by making ratings, availability, and primary actions glanceable — not just polishing chrome.

Usability findings accordion, booking before/after, and onboarding flow screen grid — Temp archive — test + onboarding evolution

Impact

Lessons

First idea ≠ right idea. Housing felt obvious; exchange mentorship was what people needed.
Cut is a design skill. Saying no to chatbot / housing / emergency kept the four-week build shippable.
Thumb zone is real. Booking only worked once it left the top bar.

Next: design for mentors — what makes them show up and stay.

Conclusion with reflections, next steps, acknowledgements, and project footer — Temp archive — conclusion