Skip to main content
  1. Projects/

Theia Sense

Table of Contents

Full-stack, microservice-based application for aesthetic image curation. Features a decoupled FastAPI backend, a dedicated ML inference service, and a responsive frontend — deployed on Azure Container Apps and Hugging Face Spaces with independent auto-scaling.

Project Link: theia.rutansh.dev

GitHub: github.com/theia-sense/theia-sense

Languages: Python

Frameworks: FastAPI, ONNX Runtime, Docker

Infrastructure: Azure Container Apps, Hugging Face Spaces

Details
#

  • An AI-powered curation model analyzes and scores user-uploaded images for aesthetic quality, using ONNX Runtime to reduce final container image size by 75% and improve inference speed
  • A dynamic ranking algorithm adapts results to the collection’s overall quality, providing a personalized user experience
  • Deployed as microservices on Azure Container Apps, enabling independent scaling of FastAPI backend and ML inference service

Related