India Enterprises use LLMs for over a billion local language speakers

India is emerging as a pivotal producer of AI across virtually every industry, driven by thousands of startups catering to its multilingual and multicultural population while also scaling to global users. 
  
India ranks among the top six global economies in generative AI adoption and has experienced explosive growth in its startup and investor ecosystem, surging to over 100,000 startups in 2023, up from fewer than 500 in 2016.

Among these, more than 2,000 startups are part of the NVIDIA Inception program, a complimentary initiative designed to foster innovation and growth. This program offers technical training, tools, go-to-market support, and opportunities to connect with venture capitalists through the Inception VC Alliance, helping startups accelerate their journey in the competitive AI landscape.

These projects are focused on developing language models for Indic languages and English, aiming to enhance customer service AI agents, rapidly translate content, and broaden access to information for a diverse population of over 1.4 billion people.

To support these initiatives, NVIDIA has launched a small language model for Hindi, the country’s most widely spoken language with over half a billion speakers. This model, named Nemotron-4-Mini-Hindi-4B, is now available as an NVIDIA NIM microservice, allowing for easy deployment on any NVIDIA GPU-accelerated system to optimize performance.

Tech Mahindra, an Indian IT services and consulting firm, is the first to utilize the Nemotron Hindi NIM microservice to create an AI model called Indus 2.0, which focuses on Hindi and its numerous dialects. Indus 2.0 leverages Tech Mahindra’s high-quality fine-tuning data to enhance model accuracy, creating opportunities for clients in sectors like banking, education, and healthcare to provide localized services.

Tech Mahindra showcased Indus 2.0 at the NVIDIA AI Summit, which is taking place from October 23-25 in Mumbai. Additionally, the company employs NVIDIA NeMo to develop its sovereign large language model platform, TeNo.

Innovators, major enterprises, and global systems integrators in India are leveraging NVIDIA NeMo to build customized language models. Companies participating in the NVIDIA Inception program for cutting-edge startups are utilizing NeMo to develop AI models for various Indic languages.