An AI-powered SAT prep platform that personalizes learning for students and simplifies progress tracking for teachers.
- Lead Product Designer
- CEO, CTO, 4 designers, and 5 engineers
AI-powered SAT prep that personalizes learning for students and simplifies progress tracking for teachers.
- Led a team of 4 designers to build Brainbox from the ground up.
- Owned the full process: concept validation, user research, prototyping, development, and testing.
- Created and maintained the design system to ensure consistency and scalability.
- Shaped the brand design to align the product identity with user needs.
The SAT went adaptive — and students had no way to prepare for it.
The SAT shifted to an adaptive format where question difficulty adjusts to performance. In Taiwan, schools don't teach SAT-specific strategies, so students rely on self-study or prep schools — where imbalanced teacher-to-student ratios make personalized guidance hard to get.
That gap pointed to a clear need: help students adapt to the new format while delivering tailored support at scale.
Use AI tutoring to give every student a personalized learning experience — whether studying independently or inside a prep school.
I studied the market and our competition and found two camps: courses with real teachers, and large practice-question banks with no teacher support. Brainbox fits the latter — but stands apart with an AI virtual tutor, something no competitor offered. I also interviewed the client, who owns a test-prep institution, to ground the work in how students actually study.
With the scope initially open, we started by identifying the features that would best serve users, then narrowed to the essentials to hit our launch deadline. We aligned through multiple rounds of stakeholder discussion — users, client, and engineers — and iterative market research.
One platform for solo learners and the classrooms around them.
- Initial assessmentA short assessment gauges the student's current performance level.
- Personalized practiceQuestions are tailored to performance, with step-by-step guidance throughout the practice journey.
- Detailed reportsAfter quizzes and mock tests, comprehensive reports highlight weak areas to practice next.
- AI tutor assistanceStudents can ask the AI tutor questions during practice and report sessions for instant help.
- Organize group practiceTeachers set up group practice sessions to monitor student performance.
- Comprehensive reportingAccess individual student reports or a consolidated group report to track progress and spot gaps.
- All individual featuresStudents in group sessions keep every individual feature, including the AI tutor.
From information architecture and user flows to low-fidelity wireframes, then high-fidelity mockups and prototypes. We ran usability and A/B tests to keep the design aligned with what students actually needed.
We handed design assets to the frontend and backend engineers, and partnered with a prompt engineer so the AI tutor felt native to the experience. Close communication kept the build true to the design and quality bar. A look at the shipped product:
A component library built for consistency and scale.
Leading a large design team taught me how much clear communication carries a product — keeping everyone aligned to the vision was what made Brainbox cohesive. Owning the whole product also sharpened my eye for design blind spots and overall quality.