01
The Problem
Beauty product discovery is overwhelming. Reviews are scattered, biased, and rarely personalized, leaving shoppers guessing and returning products that simply do not work for them.
02
What It Does
An ML-powered recommender that scrapes real review data, builds a structured ETL pipeline, and uses cosine similarity scoring to surface products that genuinely match a user's preferences and skin profile.
03
Why I Built This
As someone who spends too long researching skincare, I wanted a tool that learns from real community experience rather than sponsored recommendations. The data was out there and it just needed a proper pipeline.