Deccan AI Secures $25M to Advance AI Model Training with India-Based Expertise

The AI industry’s rapid evolution has created a surge in demand for specialized post-training services such as data generation, evaluation, and reinforcement learning. Deccan AI, founded in 2024, has positioned itself as a key provider in this space by focusing on high-quality, expert-driven work. Recently, the company raised $25 million in a Series A funding round led by A91 Partners to expand its operations.

Deccan’s business model leverages a large workforce based primarily in India, including students, domain experts, and PhDs. With about 125 employees and over a million contributors, the company ensures access to specialized knowledge necessary for the complex demands of frontier AI labs like Google DeepMind and Snowflake.

Accuracy is paramount in post-training AI tasks, as errors can significantly impact model performance. Deccan addresses this challenge by delivering precise, domain-specific data and evaluations, often under tight time constraints, balancing speed with quality.

The company is also broadening its scope beyond traditional text-based AI models to include robotics and vision systems, supported by products like the Helix evaluation suite and an operations automation platform that integrate AI models with external software tools.

Deccan’s strategic focus on India for its workforce allows for tighter quality control compared to competitors sourcing globally. While it has started sourcing niche expertise from other countries, India remains central to its operations.

For enterprises aiming to advance their AI capabilities, Deccan AI exemplifies the value of expert-led training services that combine quality, speed, and domain expertise. Engaging with such partners can be a critical step toward successful AI deployment.