At the heart of the AI machine: those who build GenAI from the ground up

A mammoth exercise is underway in the country as several organisations try to create and improvise foundational large language models (LLM) trained on vast amounts of data in local Indian languages. Firms such as New Delhi-based Centre for Studies of Tradition and Systems, Tamil Nadu-based DesiCrew Solutions, Bengaluru-based Karya, Hyderabad-based KeyPoint Technologies, Mumbai-based Navana Tech, and Noida based Devnagri have hired thousands of staff to contribute multimodal datasets–text, image, video, and voice–to train AI models.