Outfit Inspiration

I led frontend development of an AI-powered outfit builder that uses visual-similarity matching to let members swap individual pieces within stylist-curated looks, with real-time size and inventory availability. The feature takes users from browsing curated outfits to checkout in just a few taps, turning outfit inspiration into an easy path to purchase.

Contributions

  • Designed a tile-based layout and naming system to consistently display flatlay imagery across outfits and item variations.
  • Built and tested an interactive mobile prototype in Vue, validating the flow on physical devices before moving to production.
  • Implemented visual-similarity matching so members can swap individual pieces within a stylist-curated outfit while preserving the overall look.
  • Integrated real-time size and inventory availability directly into the swap and cart flow.

Challenges

  • Initial prototype testing on physical devices revealed the first version required too many screens and taps to move from browsing to cart. Redesigned around a leaner flow without losing clarity on what was in the outfit.
  • Needed a layout system flexible enough to handle a variable number of items per outfit, rather than a custom layout for every look.

Impact

  • Took the feature from interactive prototype to production in close partnership with design.
  • Sustained roughly 40% of outfit viewers favoriting an item, with about 29% of those favorites converting to cart adds, over 100+ days post-launch.
  • Established a reusable tile and layout pattern now available for future outfit-based discovery features.