Product Engineering · Startup
An AI-assisted fashion product combining digital wardrobe management, personalized outfit recommendations and garment understanding from user photos.
Context
Startup
Role
Co-Founder & CFO, Software Engineer
Year
2025
Stack

Problem
People who own plenty of clothes still struggle to plan outfits from what they already have, and most fashion apps are built to sell new items rather than help with what is already in a closet.
What I built
System
A user photographs a garment; the computer-vision pipeline isolates the clothing item from the photo and adds it to their digital wardrobe. The recommendation workflow then combines wardrobe contents with user preferences to suggest outfit combinations, surfaced back through the React frontend.
Key technical decisions
Scoped garment isolation to single-item photos before attempting full-outfit photos.
A narrower input requirement for users, in exchange for meaningfully higher extraction quality to build the rest of the product on.
Result
A working product spanning wardrobe management, garment extraction and outfit recommendation, demonstrated through a public landing page and used in investor conversations.
What I learned
How differently a feature gets scoped when you are also the one pricing it, pitching it and supporting it, not only building it.