Inspiration
We are thrilled to share our project Mimari-AI, which we developed for the AI Builder Hackathon 2026! As architects and developers, we wanted to bridge the gap between generative deep learning and parametric building analysis. Our goal was to automate complex spatial workflows and bring intelligent feature extraction right into the hands of designers. Key Highlights of our Pipeline: PyTorch & CNN Architecture: Custom convolutional neural networks optimized for spatial and structural imagery. Kaggle Ecosystem: Leveraging free GPU power for smooth model training and inference pipelines. Open Source Vision: Designed to be modular, scalable, and adaptable for various architectural styles. You can check out our code repository and notebook (mimari-ai-v1.30) on our profile. We would love to hear your thoughts, feedback, or any collaboration ideas. Let’s build the future of AI and architecture together! 🏗️✨ Feel free to drop a comment or upvote if you find it exciting!
What it does
We are thrilled to share our project Mimari-AI, which we developed for the AI Builder Hackathon 2026! As architects and developers, we wanted to bridge the gap between generative deep learning and parametric building analysis. Our goal was to automate complex spatial workflows and bring intelligent feature extraction right into the hands of designers. Key Highlights of our Pipeline: PyTorch & CNN Architecture: Custom convolutional neural networks optimized for spatial and structural imagery. Kaggle Ecosystem: Leveraging free GPU power for smooth model training and inference pipelines. Open Source Vision: Designed to be modular, scalable, and adaptable for various architectural styles. You can check out our code repository and notebook (mimari-ai-v1.30) on our profile. We would love to hear your thoughts, feedback, or any collaboration ideas. Let’s build the future of AI and architecture together! 🏗️✨ Feel free to drop a comment or upvote if you find it exciting!
How we built it
We are thrilled to share our project Mimari-AI, which we developed for the AI Builder Hackathon 2026! As architects and developers, we wanted to bridge the gap between generative deep learning and parametric building analysis. Our goal was to automate complex spatial workflows and bring intelligent feature extraction right into the hands of designers. Key Highlights of our Pipeline: PyTorch & CNN Architecture: Custom convolutional neural networks optimized for spatial and structural imagery. Kaggle Ecosystem: Leveraging free GPU power for smooth model training and inference pipelines. Open Source Vision: Designed to be modular, scalable, and adaptable for various architectural styles. You can check out our code repository and notebook (mimari-ai-v1.30) on our profile. We would love to hear your thoughts, feedback, or any collaboration ideas. Let’s build the future of AI and architecture together! 🏗️✨ Feel free to drop a comment or upvote if you find it exciting!
Challenges we ran into
PyTorch & CNN Architecture: Custom convolutional neural networks optimized for spatial and structural imagery. Kaggle Ecosystem: Leveraging free GPU power for smooth model training and inference pipelines. Open Source Vision: Designed to be modular, scalable, and adaptable for various architectural styles. You can check out our code repository and notebook (mimari-ai-v1.30) on our profile. We would love to hear
