Deccan AI, a startup specializing in post-training data and evaluation services for AI models, has successfully raised $25 million in its recent funding round. The company leverages a workforce of experts based in India to meet the growing demand for AI model refinement in the expanding market.
The Series A funding, entirely in the form of equity, was led by A91 Partners, with additional contributions from Susquehanna International Group and Prosus Ventures. Deccan AI’s strategic focus on outsourcing post-training tasks, such as data generation, evaluation, and reinforcement learning, aligns with the industry trend of ensuring the reliability of AI systems for practical applications.
While major AI research labs like OpenAI and Anthropic handle core model development internally, startups like Deccan AI play a crucial role in enhancing model capabilities and integrating them with external tools like APIs. Deccan’s services range from improving coding and agent functionalities to training AI systems to interface effectively with software systems.
By collaborating with leading AI labs on tasks such as expert feedback generation and reinforcement learning environment construction, Deccan AI caters to enterprises through its proprietary products like the evaluation suite, Helix, and an operations automation platform. The company’s clientele includes prominent names like Google DeepMind and Snowflake, reflecting its growing influence in the AI ecosystem.
Headquartered in the San Francisco Bay Area with a significant operational presence in Hyderabad, Deccan AI employs over 125 professionals and leverages a network of more than 1 million contributors, including students, domain experts, and PhDs. The startup’s agile workforce structure ensures timely project delivery and continuous innovation in AI model refinement.
Source: TechCrunch