With the increasing need for training and enhancing AI models, Deccan AI, a startup specializing in post-training data and evaluation, has successfully secured $25 million in its inaugural significant funding round. This achievement is bolstered by an expert workforce based in India.
The all-equity Series A funding round was spearheaded by A91 Partners, with support from Susquehanna International Group and Prosus Ventures.
While prominent AI laboratories such as OpenAI and Anthropic develop core models internally, there is a growing trend to outsource post-training tasks, including data generation, evaluation, and reinforcement learning. This shift is critical as organizations strive to ensure the reliability of AI systems in real-world applications. Deccan AI is positioning itself as a vital player among a new wave of startups catering to this need.
Established in October 2024, Deccan offers a range of services designed to enhance the capabilities of AI models, including refining coding skills and enabling systems to interact effectively with external tools such as application programming interfaces (APIs).
The startup collaborates with leading labs on various tasks, including the generation of expert feedback, conducting evaluations, and constructing environments for reinforcement learning. Deccan also provides enterprises with products like its evaluation suite, Helix, and an operations automation platform. As AI models begin to evolve from text-based processing to understanding complex “world models,” which encompasses physical environments like robotics and vision systems, Deccan’s offerings are adapting accordingly.
Among Deccan’s clientele are notable organizations such as Google DeepMind and Snowflake. The company has welcomed around 10 clients and maintains approximately two dozen active projects at any point in time, according to founder Rukesh Reddy in an interview.
Headquartered in the San Francisco Bay Area and featuring a significant operational workforce in Hyderabad, Deccan employs roughly 125 individuals and collaborates with a network exceeding 1 million contributors, including students, domain specialists, and PhDs. Reddy remarked to TechCrunch that between 5,000 to 10,000 contributors are typically engaged each month.
Around 10% of Deccan’s contributor network holds advanced degrees, such as master’s and PhDs; however, the proportion is higher among active participants depending on project demands, Reddy noted. The market for AI training services is experiencing rapid growth as the demand for efficient AI training solutions escalates.
Share this story