Week 2 - Product Review
Scentum: Semantic Search as a Product, Not a Feature
A review of Scentum as an AI-first product: fragrance search, vibe tags, vector spaces, voice input, and the difficulty of translating subjective taste into structure.
Abstract: Semantic search becomes product work when the user is not looking for a keyword, but for a feeling. Scentum can be read as world-building: dimensions, tags, scores, affordances, and vocabulary that let people explore. The ML engineering lesson is modest but deep: a useful coordinate system can matter as much as the model.
LinkedIn angle
Write for AI product builders who want to understand semantic search beyond a vector database checklist.