Case study
UX Design · Product Research · Mobile
A practice-first learning app that turns skill-building into modules — with a path for what to do next, and feedback when you get stuck.
Product Designer · Capstone · 2026 · Solo designer · iOS
Tools·Figma · FigJam · Google Meet
Path Learning is a project-based mobile app for self-directed learners who are tired of watching without doing. Instead of course catalogs and completion badges, the product is built around illustrated skill paths and modules — a lesson, an assignment that produces an artifact, then AI and mentor feedback. I led research, IA, wireframes, visual design, and two rounds of usability testing as sole designer.
Passive video consumption creates a false sense of mastery that disappears the moment someone opens a blank project.
Feedback is either generic praise or binary pass/fail — not category-specific enough to change lighting, composition, or logic.
Fragmented catalogs cause choice paralysis, so people spend their limited free time searching instead of practicing.
Flip the default EdTech loop from Watch → Read → Quiz to Try → Struggle → Learn → Feedback → Retry, so every concept is immediately practiced and every completed assignment becomes portfolio evidence.
The research question was how self-directed learners move from theory to practice — and where they abandon that jump. I interviewed self-taught creatives, mid-career switchers, and hobbyists, and audited Maven, Exercism, and MentorCruise to map gaps in feedback, structure, and mentorship.
Structured sessions on how people currently learn, where projects stall, and what makes feedback feel actionable versus empty.
Compared Maven, Exercism, and MentorCruise on practice loops, mentor access, and whether progress felt like portfolio work or video check-offs.
Synthesized quotes into three problem clusters — tutorial dependency, feedback quality, and choice overload — which became the product’s north star.
problem clusters from affinity mapping
competitor platforms audited
moderated usability participants
28 · Marketing Coordinator · Atlanta
“I’ve watched a hundred YouTube tutorials and I still can’t build anything from scratch. I need a path — not more content.”
24 · Photographer · Portland
“I have a hundred ideas. I just freeze the second I open a blank file. I need someone — or something — to just get me unstuck.”
35 · ML Research Scientist · Cambridge
“I don’t want to just build it. I want to understand why it works — and why every alternative approach doesn’t.”
Research ran across four weeks — planning, interviews, synthesis, then concept testing. Design work followed the same loop the product teaches: try, get feedback, iterate. Two usability rounds (Phase 1 and Phase 2) validated the lesson-and-assignment model, then tightened pricing, resources, and mentor booking.
The app is organized around a practice loop, not a course catalog. After sign up or login, learners pick interests, choose a skill, follow a module path, then work through Lesson / Assignment / Feedback — with a mentor to book when they want a human eye. Home, Paths, Projects, Mentors, and Settings live in the menu.
Early screens locked the core loop before visual design: skill choice, a photography path, a lesson with upload, and mentor booking. Testing later confirmed the path-and-assignment format; the friction lived in drafts, academic depth, and pricing transparency.
Onboarding asks what do you want to learn? after Choose Your Interests — searchable skill chips (Design, Photography, Marketing) instead of a course catalogue. Search on the interests screen is how a freeform goal still becomes a path.
Participants preferred a direct “What do you want to learn?” over a course catalogue — it felt tailored rather than like shopping.
Busy professionals wanted directed, short practice — not another hour-long lecture before they could start.
Career switchers wanted ROI and plan details early. Phase 2 added a pricing step without hurting completion.
The brand needed to feel like a studio for practice — playful illustration, not another gamified course app. A Newsreader wordmark in forest green, mint fields, and charcoal actions keep skill cards and completion states distinct without turning the product into a toy.
Mint (#E3F5DF) holds splash and sign-up so illustration can sit on a field instead of a white void. Lilac is login. Charcoal (#242527) is the only high-contrast action. Forest (#388068) is reserved for the wordmark. White is the working surface. Sage, blush, and sky mark adjacent skills without turning the product into a toy.
Paper
#FFFFFF
Buttons, search, working surfaces
Charcoal
#242527
Primary buttons, selected chips, text
Forest
#388068
Wordmark
Mint
#E3F5DF
Splash, interests, path fields
Sage
#C3E7BB
Skill-path chips
Lilac
#D7CFFE
Skill and completion accents
Blush
#F0D9F7
Secondary skill accents
Sky
#B6D8FE
Secondary UI accent
Newsreader is the Path Learning wordmark. Poppins carries H1 and H2 so “Choose Your Interests” and skill names stay friendly at phone size. Work Sans is the body — search, chips, and long lesson copy — so it doesn’t compete with the display type.
Newsreader
Wordmark
Poppins
H1 and H2 headlines
Work Sans
Body and UI
The product speaks in short actions, not catalog copy: “Get Started,” “Sign Up,” “Login,” “Choose Skills,” “Choose Skill,” “Next Lesson,” “Complete Lesson.” Feedback is a score plus notes — mentor and AI — instead of a vague “looks good.”
High-fidelity screens sit on mint, lilac, and white, with charcoal actions and line illustration. Splash offers Get Started, then Sign Up and Login; onboarding asks what you want to learn; photography opens as a Visual Storytelling path of modules; each lesson splits into Lesson, Assignment, and Feedback — including mentor and AI scores — then a live booking surface. A menu of Home, Paths, Projects, Mentors, and Settings holds the rest of the app.
Path is a practice loop of modules, assignments, and feedback — not a video library with a progress bar. Four surfaces carry the core loop; the mentor calendar sits one step up when automated notes run out.
A short lesson, then an artifact — choose a camera, upload a photo — instead of finishing a video and stopping.
Photography opens as Visual Storytelling: Framing 101 through Recreate a Film Still, so the next module is a list, not a search. The home card later shows percent complete so progress is a project, not a video checkbox.
An overall score with separate mentor and AI notes — lighting, composition, color — instead of a pass/fail.
Live availability on Sarah Chen’s calendar when the assignment needs a human eye. Testing made the calendar easy; the remaining request is a lower-cost quick critique.
Two rounds of remote moderated tests with 5 participants (career switchers, an active maker, a theoretical learner, a structured learner). Phase 1 proved the lesson-and-assignment loop. Phase 2 measured iterations on pricing, module progression, collapsible deep-dives, and mentor booking.
onboarding completion in both test rounds
mentor booking difficulty after iteration
archetypes tested across two phases
The product only works if the first session produces work, not a watched video. The path, the assignment, and the feedback ladder all exist to get someone from a blank page to an artifact they can show. What I’m still holding: a true draft/sandbox, and a cheaper asynchronous mentor option so live calls aren’t the only human door.
Designing a full practice loop — from path choice through mentor booking — and watching testers describe it as job-ready progress instead of another course to finish.
Next