Bharathi Kannan N.

I am a PhD student at the Marine Data Science Lab, University of Rostock, Germany. My research focuses on Generative AI for Building Information Modeling (BIM), where I work on the Show2Instruct project.

Previously, I completed my M.Sc. in Computer Science (Visual Computing) at the University of Siegen, where my thesis explored depth fusion techniques for transformer-based video object detection.

Email / GitHub / LinkedIn / Google Scholar / X (Twitter)

Research Interests

My research lies at the intersection of Large Language Models and multimodal AI, with a particular focus on tool-augmented LLMs, multimodal spatial reasoning, and generative methods for BIM/CAD automation.

Publications

MCP4IFC: IFC-Based Building Design Using Large Language Models
Bharathi Kannan Nithyanantham, Tobias Sesterhenn, Ashwin Nedungadi, Sergio Peral Garijo, Janis Zenkner, Christian Bartelt, Stefan Lüdtke
arXiv Preprint, 2025

An open-source framework enabling LLMs to create, edit, and query IFC building models through the Model Context Protocol, combining predefined BIM tools with dynamic code generation via retrieval-augmented generation.

paper / code / project