Tech Mahindra has launched an 8-billion parameter Hindi-first education LLM under the IndiaAI Mission at the India AI Impact Summit in Delhi, becoming the only IT services company selected by the mission to build its own foundational LLM.
The new model is an advanced version of the company’s earlier 1.2-billion parameter foundation model built under Project Indus. The Pune-headquartered firm scaled it up to an 8-billion parameter architecture, positioning it as a sovereign, education-focused LLM designed for Indian linguistic and cultural contexts.
The model has been developed in collaboration with NVIDIA and is powered by NVIDIA AI Enterprise. It uses the NVIDIA NeMo framework for training and NVIDIA NIM microservices for deployment to ensure scalability and production readiness.
To tackle data scarcity in certain Indian languages, the company generated 500 million synthetic tokens using NVIDIA NeMo Data Designer.
The LLM will debut as an education-first model aimed at democratising access to high-quality learning. It is designed to support foundational understanding in subjects such as physics and other disciplines, while operating fluently in natural Hindi.
The company said the model also enables agentic AI capabilities, allowing the creation of autonomous AI agents that can interact in Hindi for learning use cases.
Nikhil Malhotra, Chief Innovation Officer and Global Head of AI and Emerging Technologies at Tech Mahindra, said global foundation models are often not built for countries with deep linguistic 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,” Malhotra said.
John Fanelli, Vice President, Enterprise Software at NVIDIA, said the demand for sovereign AI is accelerating globally. “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.”
Project Indus is part of India’s broader sovereign AI push, with a focus on scalable models grounded in local languages and regulatory needs. With the 8-billion parameter upgrade, Tech Mahindra is positioning the education LLM as a one-stop platform for academic use cases, aligned with national priorities and the IndiaAI Mission’s objectives.
The company said it is also exploring advancements from NVIDIA Nemotron, NVIDIA’s open model and research ecosystem, to accelerate future sovereign model development and improve instructional quality for India-specific education and public-sector applications.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.