PeerUp
Connecting Students for Collaborative Learning
Think Bumble for studying, a peer learning platform where AI matches students by academic goals and availability, so finding a study partner is as seamless as finding a date.
Connecting Students for Collaborative Learning
Think Bumble for studying, a peer learning platform where AI matches students by academic goals and availability, so finding a study partner is as seamless as finding a date.
Type: Solo iteration on a group academic project
Role: Product Designer & Manager
Timeline: 1 month (redesigned to leverage AI matching)
Tools Used: Figma, Adobe XD, User Interviews, Persona and Scenario Development, UX Research, Competative Analysis, User Testing, Wireframe & Prototype
Objective: Address the growing need for personalized, peer-based support systems in education, especially in hybrid and remote learning environments. PeerUp is designed to connect students based on academic goals, learning styles, and complementary strengths to foster more effective study partnerships.
User interviews revealed that students knew who they wanted to study with but had no frictionless way to coordinate. With the rise of hybrid and remote education, that same friction got worse. Research confirms that peer supported learning improves retention, understanding, and motivation, yet no purpose built tool existed to make peer finding effortless. The dating app market proved AI powered matching works at scale. Students have the exact same coordination problem, just applied to academics.
The design hypothesis: if we reduce scheduling friction to near zero and surface compatibility signals clearly, peer learning becomes a default behavior rather than an effortful one. By incorporating learning theories like Constructivism and Social Learning Theory, the platform encourages users to teach and learn from one another, reinforcing their own knowledge while benefiting others.
Designed primarily for high school and college students who feel isolated in remote learning environments, lack access to traditional tutoring, or find it hard to seek help from TAs and professors. The platform also serves educational institutions that want to boost student engagement and retention through peer learning opportunities.
The original version used a manual matching model where students posted their availability and needs. User testing revealed that manual effort killed adoption: if pairing was not automatic, it did not happen.
The redesign was driven by the rise of AI and a key market insight: dating apps had already proven that AI powered matching at scale changes user behavior. Students have the same coordination problem. Two shifts drove the new product direction: automated matching replaced manual posting, and schedule compatibility was surfaced upfront rather than treated as a separate step.
Reduce Onboarding Friction: Design social mechanics that make the first peer match feel immediate and rewarding, not effortful.
Retention and Impact: Measure success through user retention rates and educational outcomes such as improved grades or user reported comprehension.
Personalization: Build a recommendation interface that makes the right peer feel obvious, using adaptive matching based on learning styles, skill gaps, and academic goals.
1. Reducing Onboarding Friction
Challenge: Students are reluctant to adopt yet another platform if it does not demonstrate value immediately.
Decision: Prioritized immediate peer matching as the first action in onboarding, paired with small rewards to create an early win. Delayed profile completion to after the first match to reduce upfront effort.
2. Designing for Diverse Learners
Challenge: High school and university students have different coordination norms and academic structures.
Decision: Built customizable institution level settings so the platform could be tailored to specific curricula, rather than trying to build a single universal experience.
3. Measuring Learning Outcomes
Challenge: Peer collaboration impact is hard to attribute directly to a platform.
Decision: Implemented pre and post session data collection and surfaced personal analytics to users, making the value visible rather than asking them to trust it.
With the rise of AI, the original manual pairing model became obsolete. I iterated on the idea and refocused the technology to be more automated and personalized, eliminating the friction of manual scheduling entirely.
Below is the process book from the original ideation, research, and prototype phase.