TAMIL-TO-CODE: LEVERAGING LARGE LANGUAGE MODELS TO TRANSLATE INDIGENOUS KNOWLEDGE INTO FORMAL COMPUTATIONAL REPRESENTATIONS

The resurgence of interest in preserving and revitalizing indigenous languages has found a powerful ally in artificial intelligence. This paper proposes a multimodal AI framework for translating ancient Tamil texts into structured symbolic representations suitable for applications in policy modeling, healthcare decision support, and educational content structuring. By integrating large language models (LLMs) with symbolic reasoning systems and domain-specific ontologies, this framework aims to computationally encode centuries of indigenous Tamil knowledge while maintaining its cultural integrity and linguistic nuances.