Adarsh Singh
All work

Fomofy

A fashion app where every swipe teaches the feed what to show next.

Role
Led the backend, as CTO of the team that built it
When
2025 to now
Built with
TypeScript, Express, PostgreSQL, Redis, Flutter
Code
Private repository

Linen shirt

Sand, linen

  • Earth
  • Linen
  • Relaxed
  • Brunch

What the feed has learned

Nothing yet. Drag the card, or use the buttons. Each choice changes which card comes next.

The code is private. This is a working sketch of the idea.
Live
on iOS and Android, built for Kreative Motion Ventures
6
style signals behind each feed: trends, size, body type, comfort, occasion, skin tone
Fewer
colour-related returns, with add-to-cart conversion up

What it is

Fomofy is a fashion shopping app for iOS and Android, built by our team for Kreative Motion Ventures. Generic feeds ignored size, body type, and taste, so discovery felt like noise. Fomofy builds the feed from each shopper’s own signals instead, and shows curated homegrown labels one piece at a time. Keep it or pass.

Every swipe is a signal

The feed learns from what you keep and what you pass. That only works if every product is described the same way the shopper is, so each piece carries its fit, fabric, colour family, and occasion, and the ranking compares like with like.

Colour and fit

Skin-tone matching detects tone and undertone, maps it to a palette, and flags which colours complement or clash, which powers a “best for you” filter. Virtual try-on overlays a garment on a photo the shopper uploads. Both went after the most common reason for a return.

The commerce underneath

Brands bring their catalogues in through Shopify. Orders, shipping across several couriers, returns, and brand payouts each sit behind a single interface, so adding a courier or a payment rule does not mean touching order code.

Elsewhere