Mismatched expectations
Mentees often don't know how to describe what they need. Mentors struggle to clearly present what they offer. So the first message rarely lands.
Case Study · 0 → 1 · Founding Designer
A path to better mentorship for early-career STEMs.
PM & founding designer for MentorUp — a 0→1 platform built to match mentees with the right mentors, then keep them company through the emotional gap between sessions. A community for job seekers and early-career designers when the search gets long.
Visit MentorUp ↗
@Mia · Mentor
"You don't have to figure it out alone — let's break it down together."
AI Mentor
"Hey, late-night thought spiral? Here's one small thing to do tomorrow."
Background & my role
- Product
- Role
- Scope
- Outcome
- Live
01 / The brief
No product. No flow. Only a vision.
When I joined MentorUp, there was nothing on the screen — only a sentence the founders kept repeating: "help early-career designers and job seekers find meaningful mentorship."
As the first and only designer, I owned everything from defining the MVP, mapping user journeys, building the design system, to shaping the 0→1 product experience. I worked closely with the founders as the bridge between business goals, user needs, and technical feasibility — setting the foundations for a product that could actually launch.
02 / Problem Space
Most mentorship platforms fail twice.
Users can't find the right mentor — and they don't trust what they see. From early interviews, three frustrations kept showing up.
Trust gap
Users hesitate to book a session because the platform doesn't communicate credibility or personality. Static cards don't tell you who someone really is.
Emotional needs ignored
For juniors and job seekers, mentorship isn't only about knowledge — it's about encouragement, confidence, and someone "on their side" between sessions.
03 / Why this matters
The job hunt isn't a search problem. It's a confidence problem.
"I've sent 200 applications. I've stopped opening the rejection emails. I keep wondering if the problem is me."
— interview, junior designer, 8 months job-huntingBehind every "looking for mentorship" message we read, there was a quieter sentence: I'm losing trust in myself.
For early-career STEMs, the climb is rarely linear. Portfolios get reworked at midnight. Recruiter silence stretches into weeks. Friends move on; doubts move in. Most platforms hand these people a search bar and call it a day.
MentorUp had to be more than a marketplace. It had to be a place where someone — human or AI — was always ready to say, "I see you. Here's the next small step." A community of values, not just a directory of profiles.
-
01
Belonging over browsing
People don't open MentorUp to shop. They open it because they want to feel less alone in the process.
-
02
Continuity over one-off calls
One great session doesn't fix burnout. The product had to live in the spaces between sessions.
-
03
Encouragement as feature
Confidence is the real deliverable. Every flow had to leave the user a little more sure of themselves.
04 / Problem Statement
How might we design a mentorship platform that not only matches mentees with the right mentors, but also builds trust and provides continuous emotional support throughout their job-seeking journey?
05 / MVP Flow
How MentorUp supports you, step by step.
After defining the problem space, I translated the insights into a focused MVP — a clear, end-to-end path from "I need help" to "we just had a session" — designed so people felt held, not sorted, at every step.
-
01 · Browse mentors
Filter by domain, stage, and tone. Profiles surface personality first, credentials second — so the human shows up before the resume does.
-
02 · Articulate what you need
The platform helps mentees translate vague struggles ("I feel stuck") into a clear ask — so the first message lands and the session starts in the right place.
-
03 · Build trust before booking
Every mentor profile shows real session reflections, response style, and what they actually focus on. Trust is earned in scrollable detail, not stars.
-
04 · Book, meet, reflect
One-flow booking + a post-session reflection prompt that's fed back into the next match. Each session leaves a fingerprint on the next one.
06 / Supporting flows
Even a "simple" MVP runs on invisible flows.
As the first designer, I quickly realized the four-step main journey only worked because of dozens of supporting flows behind it — the quiet plumbing that makes a product feel coherent end-to-end.
Search & filter
Combinable filters, empty states, "no results" recovery, and recently-viewed memory.
Booking & confirmation
Time-zone handling, reschedules, payment edge cases, and confirmation states that reassure rather than confuse.
Mentor onboarding
Profile review, calendar sync, payout setup, and a tone-of-voice prompt that shaped how mentors introduced themselves.
Post-session reflection
Lightweight notes, follow-up nudges, and a feedback loop that fed the next match — so progress kept compounding.
07 / Design system
A system that scales with the team.
With the core journey defined, the next challenge was ensuring the product could grow with clarity and consistency. So I built MentorUp's design system from the ground up — tokens, components, patterns, and writing principles — small enough to ship fast, structured enough to outlive any single screen.
- Tokens. Color, type, spacing, radius, elevation — one source of truth.
- Components. Buttons, profile cards, filter chips, message bubbles, empty states.
- Patterns. Booking flow, onboarding, notification states, mentor/mentee dual views.
- Voice. Warm, plain-spoken, encouraging — never corporate.
08 / Dual user experiences
One product. Two minds.
After defining the system, the next step was applying it consistently across the product — for both mentees and mentors, whose needs are wildly different. Easy to spot by one tiny CTA: "Become a Mentor."
Mentee view
- Discovery, comparison, trust-building.
- "Become a Mentor" CTA visible at the top.
- Personality-first profile cards.
- Emotional copy: "Find someone who gets it."
Mentor view
- Schedule, requests, earnings, growth.
- No "Become a Mentor" CTA — they already are.
- Operational density, calmer tone.
- Reflection prompts to deepen mentor impact.
09 / Post-launch insights
Once the MVP launched, the data surprised us.
68%
Came in from mobile
About 68% of early users opened MentorUp on their phones — direct from social marketing, never via desktop. Mobile responsiveness became an immediate roadmap priority.
7/7
Engaged in stress
All seven interviewed users mentioned opening the app during stressful or uncertain moments — between applications, after rejections, late at night. We were a companion app, not a marketplace.
10 / Responsive mobile design
The phone became the primary canvas.
When 68% of users walked in through a phone, "responsive" stopped being a polish task. Every flow we'd designed for desktop had to feel native on mobile — not shrunk, not stacked, but reconsidered for thumbs and small moments.
- Re-laid every primary screen for portrait-first reading.
- Bottom-anchored CTAs, single-tap booking, gesture-friendly filters.
- Tightened the AI mentor surface so it felt like a chat, not a form.
- Optimized profile media so trust still built in 4G conditions.
11 / Emotional context
People weren't opening MentorUp to browse. They were opening it because something hurt.
Between job applications, late-night doubts, interview rejections, and moments of deep uncertainty, many mentees felt alone. Human mentors were incredibly helpful — but they couldn't be there 24/7. And that emotional gap between sessions was where anxiety grew the fastest.
So we reimagined what "mentorship" could mean. Instead of building just a smarter FAQ, we built something more human — a layer that could hold space at 1 a.m. when no human was awake to.
12 / The AI mentorship layer
Five surfaces. One companion.
Each surface meets the user where the doubt actually shows up — before, during, between, and after sessions.
FAQ Chatbot
Instant answers to the everyday questions — pricing, scheduling, how to prep for a first session — so humans aren't blocked at the door.
AI Mentor
The 24/7 companion. Listens, reframes, suggests one small next step. Tone adapts to whether the user is anxious, exhausted, or hopeful.
AI Consulting
Resume read-throughs, interview rehearsal, portfolio gut-checks — practical strategy work that doesn't always need a paid session.
Just-in-Time Assistance
Surfaces help at the exact moment of friction — drafting an outreach DM, untangling a feedback note — without making the user ask.
Behavior-Driven Tooltip
Watches for hesitation patterns (re-reading a profile three times, abandoning the booking flow) and offers a gentle nudge — not a popup.
13 / AI personalization · in action
Same question. Different person. Different tone.
Every user comes with a different story, so the AI adapts to theirs. A junior doubting their abilities, a new grad in interview cycles, a long-term seeker fighting burnout — each gets a different tone, a different style of encouragement, a different next step. Click a persona — the AI mentor rewrites itself.
""
Concept prototype · models the persona-aware tone-shift logic the live MentorUp AI mentor uses.
14 / Results
Launched as a 0→1 mentorship platform — both layers proven.
MentorUp successfully launched, validating both the core mentorship flow and the extended AI-supported experience. Within the early launch phase, the product onboarded 500+ early-stage users — confirming a real demand for accessible, guidance-driven mentorship.
0→1
Built from nothing
500+
Early-stage users onboarded
2/2
Layers validated · core flow + AI
15 / Reflections
Being the first and only designer meant designing more than screens.
I shaped the product logic, the design system, and the experience strategy from the ground up. The biggest lesson wasn't about pixels — it was about learning how to make thoughtful tradeoffs, moving fast without sacrificing empathy.
- —Even an MVP requires many invisible supporting flows to truly function.
- —Trust and emotional context matter as much as usability in mentorship products.
- —AI is most effective when it complements human connection rather than replacing it.
- —Community is a feature. Belonging is the deliverable.
"If given more time, future iterations would focus on deeper personalization and long-term support — turning the platform from a place you visit into a community you belong to."