Mercor competitor Deccan AI raises $25M, sources experts from India

Deccan AI, a startup that supplies post-training data and evaluation work to frontier AI labs, has secured a significant $25 million in its first major funding round. This development is noteworthy as it highlights the growing demand for specialized services in AI model refinement and training. The all-equity Series A round was led by A91 Partners, with participation from Susquehanna International Group and Prosus Ventures.

The increasing complexity of AI models has led to a shift in the way companies approach their development. While some frontier labs like OpenAI and Anthropic build core models in-house, they often outsource post-training work to specialized startups like Deccan. This work includes tasks such as data generation, evaluation, and reinforcement learning, which are essential for making AI systems reliable in real-world use.

Deccan provides a range of services, from helping models improve coding and agent capabilities to training systems to interact with external tools like APIs. The startup works with frontier labs on tasks such as generating expert feedback and running evaluations, while also serving enterprises through products like its evaluation suite, Helix, and an operations automation platform.

Deccan's customers include Google DeepMind and Snowflake, and the startup has onboarded about 10 customers and runs a couple of dozen active projects at any given time. With a large operations team in Hyderabad and a network of over 1 million contributors, Deccan employs about 125 people. Around 5,000 to 10,000 contributors are active in a typical month.