Quizlet's content discovery team reconceived the recommendations experience around a a high-leverage insight: students studying for a specific course don't need more content choices. They need the right content curated for that course, without having to dig aimlessly or create their own materials manually.

I scaled initial findings on course-focused discovery into a v1.0 which made that insight pervasive: clear course-to-content associations across recommendation surfaces, user-created and Quizlet-curated course folders containing top peer-vetted content recommendations, and a retooled recommendations model that powered it all through a hybrid approach combining machine learning ranking with manual content tagging.