Bengaluru’s CognitiveLab Bags Meta Grant for Multilingual AI!
In brief · AI summary Meta has awarded its 2024 Llama Impact Grant to CognitiveLab, a Bengaluru-based AI research lab, for Nayana, a multilingual and multimodal AI system that uses open-source Llama models to process documents, images, and speech in over 22 global languages. With the grant, the lab aims to expand language support, improve multimodal integration, release new datasets, and optimise deployment tools for low-resource areas. CognitiveLab's founder, Shashi Kumar, said the grant will accelerate the goal of building AI for underrepresented languages and communities.
Meta has awarded its prestigious 2024 Llama Impact Grant to CognitiveLab, a Bengaluru-based AI research lab, for its groundbreaking work on Nayana, a multilingual and multimodal AI system designed to serve linguistically diverse populations.
Nayana: making AI speak more than just english
CognitiveLab’s Nayana project uses open-source Llama models to process documents, images, and speech in over 22 global languages, including several low-resource Indic languages. The system has already surpassed industry standards in optical character recognition (OCR) for 10 Indian languages.
With the Llama Impact Grant, the lab aims to:
- Expand language support
- Improve multimodal integration (text, speech, and vision)
- Release new datasets
- Optimise deployment tools for low-resource areas
Building Open-Source AI for Billions
Founded in Bengaluru, CognitiveLab is known for initiatives like the Indic LLM Leaderboard and public datasets that aid global research. Its mission is to make AI inclusive, transparent, and community-driven.
CognitiveLab’s founder, Shashi Kumar, noted the grant will accelerate their vision to build AI that works for underrepresented languages and communities. He credits open-source models like Llama 4 for making high-impact development possible with minimal resources.
Meta’s support for Open AI
The Llama Impact Grants support projects using Meta’s open-source LLMs—Llama 2, 3, and 4—for social good. Meta India’s Public Policy Head, Shivnath Thukral, highlighted Nayana as a shining example of how open AI bridges digital divides in multilingual countries like India.
India at the Forefront of Llama Adoption
With over 1 billion global downloads and 85,000+ derivative models, Llama has become a key player in AI innovation. India is now one of the top adopters of this technology, with organisations like CognitiveLab leading the charge in applying it to real-world challenges.
As Nayana evolves, it promises to bring AI accessibility to more than 3 billion people, empowering communities that have long been left out of the digital revolution.