Deccan AI raises $25M
Deccan AI closed a $25M round to scale India‑based annotation and AI training operations, signaling continued VC appetite for high-quality data labeling supply chains that feed models. Investors see onshore training workforces as a strategic asset for cost and quality in model development. (techcrunch.com)
The Series A was led by growth equity firm A91 Partners with participation from Susquehanna International Group (SIG) and existing backer Prosus Ventures. (TechCrunch: ) Deccan AI was founded in October 2024 by Rukesh Reddy and is headquartered in the San Francisco Bay Area while running a large operations team out of Hyderabad. (TechCrunch: ) The company sells an evaluation suite named Helix and an operations automation platform, and it runs projects to improve models’ coding and agent capabilities, integrate models with external APIs, and build reinforcement‑learning environments. (TechCrunch: ) Clients named by the company include Google DeepMind and Snowflake, and founder Rukesh Reddy said Deccan has onboarded about 10 customers while running a couple dozen active projects at any given time. (TechCrunch: ) Deccan employs roughly 125 people and reports a contributor network exceeding 1 million individuals, with approximately 5,000–10,000 contributors active in a typical month and about 10% of the base holding master’s degrees or PhDs. (TechCrunch: ) The startup is being cast as a rival to data‑and‑evaluation vendors such as Scale AI, Surge AI, Turing and Mercor as demand for post‑training data, evaluation and RLHF services grows. (TechCrunch: ) Reporting in Indian outlets says the new capital will be deployed to expand enterprise‑facing products and build higher‑accuracy systems, and local coverage notes plans to deepen on‑the‑ground presence with a Bengaluru office focused on enterprise growth. (Economic Times: Entrackr: )
