Complexity → Clarity
I turn messy data into clear flows.
The goal is understanding, not information overload.
Real Transformations
BEFORE (Complex)
10+ dimensions of question classification (difficulty, knowledge points, question type, common mistakes...)
AFTER (Clear)
"Here's the next best question for you"
BEFORE (Complex)
User says "I want to transfer money"
AFTER (Clear)
Understanding real need: "I'm not sure how to use the app"
BEFORE (Complex)
3000+ raw question data
AFTER (Clear)
Structured knowledge graph + ROI ranking algorithm
My Design Principles
Hide complexity, surface simplicity
Users don't need to know how complex things are behind the scenes. They just need a simple answer.
Progressive disclosure
Give the conclusion first, expand only if they want more. Just like this website's design.
Context-aware responses
Same question, different user backgrounds — different responses needed.
Design for the frustrated user
If it works for someone who's confused, it'll work for everyone.
Case Study: VectorPaths
How I simplified adaptive learning:
Collected 10+ dimensions per question
Difficulty, knowledge points, question type, common mistakes, time estimate, related concepts...
Built ROI calculation algorithm
Based on user's target score and current level, calculate the "return on investment" for each question
Surfaced one simple output
Users only see: "This is the question you should practice right now"
"The complexity lives in the system. The simplicity lives in the experience."
"Good design is invisible."