Daft-Daytona Hackathon - First Prize
International Hackathon
Built an AI agent for interior design using Gemini nano Banana and Nano VLM, enabling layout-preserving redesigns; won First Prize.
The Story
At the Daft-Daytona Hackathon 2025, our team tackled the challenge of AI-powered interior design with a unique approach. We built an intelligent agent that could understand room layouts and suggest redesigns while preserving the fundamental structure and flow of the space.
The core innovation was using Gemini Nano Banana combined with Nano VLM (Vision-Language Model) to process room images and generate contextually aware design suggestions. The system could:
- •Analyze existing room layouts and identify furniture, fixtures, and spatial relationships
- •Generate redesign proposals that maintain the room's usability and flow
- •Provide multiple style options (modern, minimalist, bohemian, etc.) while keeping the layout intact
- •Offer budget-conscious alternatives by suggesting similar furniture pieces
What made our solution stand out was the emphasis on "layout-preserving" redesigns—rather than generating completely new room configurations, our agent understood that users often want to refresh their space without major structural changes. This practical approach resonated with the judges and earned us the First Prize.
Technologies & Skills
Gallery
