Case study
Text and image search UI
Search UI for a NECTEC image-retrieval model: query by text, by uploaded images, or both, and see ranked results with correlation scores.
FIG.1Freelance
- Role
- Freelance frontend developer for NECTEC (2024) — 16 of 18 commits
- Stack
- ReactViteTailwind CSSDocker
- Result
- The model could be demoed and checked without calling the API by hand.
Problem
Researchers needed a usable front end to demo and test the retrieval model, and to replay past queries.
What I owned
Freelance frontend developer for NECTEC (2024) — 16 of 18 commits
What I built
- Built combined text and multi-image search with correlation scores.
- Added replay of API results with download of the query and timing.
- Packaged the app for Docker and fixed mobile layouts.
System shape
React + Vite + Tailwind client calling the model API through a proxy, with a health indicator, error handling for uploads, and Docker builds.
What changed
The model could be demoed and checked without calling the API by hand.