01 — Applied AI product

Case study — garment intelligence

FitFabric

Designing an AI-assisted system for garment understanding and fit reasoning.

FitFabric explores how garment evidence, body measurements and fit preferences can produce a more understandable recommendation without hiding uncertainty.

Role

Product direction, garment reasoning, interface and prototype

Platform

iOS + API

Status

Working research prototype

Focus

Fit decision and garment understanding

FitFabric app home screen showing an active profile, fit check entry point and recent recommendations
01 / 05 — Personal profile

A profile you create once.

Store body measurements and fit preferences once, then reuse the same profile across products.

Body measurements · fit preferences
02 / 05 — Start a fit check

Start with the product.

Begin from a product image, link or size chart. Manual input remains available when the page cannot provide enough evidence.

Image · link · size chart
03 / 05 — Gather evidence

Turn scattered information into evidence.

FitFabric organizes garment details, size-chart information and unresolved questions before making a recommendation.

Demo footage coming next
04 / 05 — Explain the result

A recommendation with reasons.

The result explains the suggested size, relevant garment evidence, possible fit risks and what remains unknown.

Result footage coming next
05 / 05 — Behind the result

Evidence before confidence.

Body profile, garment evidence and size-chart information come together as confirmed facts, observations, inferences and unknowns.

Body profile Garment evidence Size chart
Recommendation / Risks / Unknowns

Why FitFabric works

Built to explain,
not just to answer.

01

Reusable profile

Create measurements and fit preferences once, then reuse them across products.

02

Less manual comparison

Organize product and size-chart evidence without requiring users to interpret every chart themselves.

03

Explainable recommendation

Show not only a suggested size, but also the evidence and fit reasoning behind it.

04

Honest uncertainty

Separate confirmed information, observation, inference and unknowns. The final decision remains with the wearer.

What I contributed

Six things,
end to end.

  • Product direction and scope
  • Garment reasoning model
  • User flow and interface design
  • AI-assisted evidence workflow
  • iOS and API prototype
  • Testing and verification