New Delhi

Tech Mahindra, a global provider of technology consulting and digital solutions, partnered with NVIDIA to launch a new education-focused Large Language Model (LLM) under its Project Indus initiative. Tech Mahindra confirmed that the Hindi-first model is built with authentic Indian linguistic and cultural context to provide sovereign and scalable artificial intelligence.

The new LLM debuting as an educational model is aimed at democratising high-quality learning across the country. It is designed to help millions of students gain a deeper foundational understanding of subjects such as physics and other disciplines while embracing the linguistic and cultural diversity of India. With support from multiple partners, the platform is intended to serve as a comprehensive resource for educational use cases.

John Fanelli, Vice President, Enterprise Software, NVIDIA, said, "The global push for sovereign AI is accelerating demand for foundation models tailored to local languages and cultural contexts. By leveraging NVIDIA AI Enterprise, Tech Mahindra delivers the production-ready performance, reliability and scale required to power Project Indus."

The development of the education LLM utilised advanced AI frameworks and infrastructure from NVIDIA, including the NVIDIA NeMo framework and deployment through NVIDIA NIM microservices. These tools enabled the model to achieve production readiness and scalability.

To manage data scarcity in specific languages, the Tech Mahindra team used innovative techniques to generate half a billion synthetic tokens with NVIDIA NeMo Data Designer. The model also supports Agentic AI, allowing for the creation of autonomous agents that communicate fluently in "natural Hindi" to facilitate AI-driven learning.

Nikhil Malhotra, Chief Innovation Officer & Global Head of AI and Emerging Technologies, Tech Mahindra, said, "AI is becoming central to national digital infrastructure and inclusive growth, but global foundational models are often not designed for countries with deep linguistic and cultural diversity like India. A key industry challenge is the lack of domain-trained language models grounded in local languages and learning contexts, particularly in education. Through Project Indus, our collaboration with NVIDIA directly addresses this gap by delivering a Hindi-first, sovereign AI model that enables scalable, relevant, and accessible AI-powered learning and citizen-centric services for India."

READ MORE: Uttarakhand: Rohit Prakash pleads for Muslim keeping shops open till late

As the project evolved, Tech Mahindra explored advancements from the NVIDIA Nemotron ecosystem to accelerate the development of sovereign models. These future enhancements are expected to shorten training cycles and improve the quality of instruction for India-specific education and public-sector needs.

Building on an earlier foundational model with 1.2 billion parameters, the company scaled Project Indus to an 8-billion-parameter architecture, which represents a significant increase in model capability.