A self-directed behavioral design reference, built and deployed
86 topics. One coherent sequence. Built in six weeks.
The best material on behavioral psychology and UX was scattered across books, conference talks, and academic papers — with no structured path connecting them. So I built one, in collaboration with Claude, which acted as curriculum co-author: mapping conceptual dependencies between topics, stress-testing module sequencing, and drafting quiz questions that actually test comprehension rather than trivia.
Organizing 86 topics into a coherent sequence required mapping conceptual dependencies — not grouping by popularity or alphabetizing, but understanding which ideas are load-bearing for others. A topic introduced before a learner has the prior context creates gaps that undermine everything after it.
I worked through the module order with Claude iteratively, treating it like a design critique: proposing a sequence, identifying where a learner would hit a gap, restructuring until the progression felt earned. The final order — human decision-making → behavior change → product understanding → design → measurement — reflects how understanding actually compounds.
A course that's just articles is a blog. Each topic needed a mechanism to create a moment of active recall — something that moves a learner from recognizing information to retrieving it. Every topic has a 4-question quiz, tracking read and quiz-passed states independently. The distinction between having read something and understanding it is made explicit.
media: /images/ux/lesson.png
Claude drafted the initial question set for all 86 topics — questions targeted at the specific misconceptions a designer is likely to carry into each concept, not generic multiple choice. I reviewed and edited every set. 344 questions, authored collaboratively.
media: /images/ux/quiz.png
The direction I wanted to pursue but didn't fully build: rather than fixed questions, the model generates examples relevant to whatever product domain a user is actually working in. Fogg's Behavior Model framed around a fitness app for one learner, enterprise SaaS for another. The data shape already supports it. The gap is product decision, not technical limitation.
The course is deployed at ux-design-course.vercel.app with 86 topics, full quiz mechanics, and progress persistence. Most portfolio pieces show what you can design. This one shows how I think about the field — and how I work. Using Claude as a collaborator throughout, from curriculum sequencing to question writing to architecture decisions, is the same approach I bring to product design. The tools are new; the underlying practice of thinking carefully before building hasn't changed.
next: /agent/jhsu-portfolio