Swadeshi: Local to Global

Professional
← Back to Projects

A Generative AI web application designed to reimagine luxury western apparel using traditional Indian textiles and art forms, blending cultural heritage with modern fashion.

Project Details

  • Role: AI Integration and Prompt Engineering
  • Platform: Web Application
  • Client: Vodafone Idea

Developed for a high-profile exhibition, Swadeshi allowed users to interactively “redesign” fashion. Users selected their favorite western garments and paired them with traditional Indian textiles (like Ikat or Banarasi). The system then used a custom-tuned AI model to generate a unique, high-fidelity image of the reimagined apparel, which users could download and share via social media.

Following are some results from the application,

Technical Contributions

  • Developed a custom-defined prompt architecture fine-tuned for the Google Nano Banana model. This included defining complex JSON objects for user-defined inputs to ensure the AI interpreted fashion textures and garment shapes accurately.
  • Engineered the beta version of the entire backend workflow using Node-RED. This allowed for a fast, visual validation of the logic before handing it over to the web development team for final production implementation.
  • Conducted deep research into various Generative AI APIs to determine the most efficient model for real-time image generation within a high-traffic exhibition environment.
  • Successfully learned, prototyped, and implemented the GenAI APIs and backend logic within an intensive 10-day development sprint.

Personal Notes

This project was an exciting departure from my usual game engine work, pushing me to learn the nuances of Generative AI and Prompt Engineering in a short timeframe. It taught me that regardless of the medium—whether it’s Games or AI, the core challenge is always the same: how to take a powerful, complex technology and make it feel like magic for the end user.

Technologies Used

Gemini Nano Banana Node-RED Python