Sarvam, an AI company, has been deploying its models in various enterprise and government applications, such as voice AI and language technologies. Co-founder Pratyush Kumar revealed that the company achieved this by training multiple models internally, avoiding reliance on external technology.
This approach highlights Sarvam’s commitment to self-sufficiency and showcases the efficiency of its model training process. By utilizing just 40 researchers, Sarvam’s ‘Frugal’ AI methodology, as described by Kumar, successfully conducted a ‘DeepSeek,’ demonstrating the company’s capability to innovate with limited external resources.
With the increasing demand for AI solutions across sectors, Sarvam’s streamlined approach to model development could set a new standard in the industry. This cost-effective and internally-driven method underscores the company’s technical capabilities and positions Sarvam as a potential player in the AI landscape.
Source: Tech-Economic Times