Deccan AI

Deccan AI

Deccan AI is an enterprise data services company that supplies expert-built training data for large language models and agents. Deccan AI provides supervised fine-tuning (SFT) and RLHF datasets, reinforcement-learning environments, model evaluations, and agentic solutions, sourcing much of its skilled workforce from India. Founded in 2023, Deccan AI focuses on the higher-skill end of the data market: coding, math, science, agentic tasks, and robotics, where domain experts and PhDs review the work. It serves frontier labs and enterprises that need accurate post-training data rather than commodity labeling, positioning against vendors like Scale AI, Surge AI, and Mercor.

Production credibility: Founded in 2023 by Rukesh Reddy; headquartered in the San Francisco Bay Area with a large operations team in Hyderabad, India. Raised a $25M all-equity Series A in March 2026 led by A91 Partners, with participation from Susquehanna International Group (SIG) and existing investor Prosus Ventures. Named customers include Google and Snowflake. The company reports about 125 employees and a contributor network exceeding one million, with roughly 5,000-10,000 active in a typical month and about 10% holding advanced degrees; contributor pay reportedly ranges from about $10 to $700 per hour. Deccan says it grew roughly 10x over the prior year to a double-digit million-dollar revenue run rate. Its product line includes STARK RL environments, Helix evaluations, and EnterpriseOS agents.

Key Features

• Supervised fine-tuning (SFT) and RLHF datasets built by vetted domain experts
• STARK reinforcement-learning environments where models train against coded, verifiable tasks
• Helix evaluation suite combining automated and human review with custom rubrics
• EnterpriseOS agents for back-office automation with humans kept in the loop
• Coverage across coding, math, agentic, multimodal, model-alignment, and robotics data
• A large India-based expert network including masters and PhD-level contributors
• Multi-step verification and quality playbooks aimed at accuracy on hard tasks
• Custom data programs scoped for frontier labs and enterprise model teams

Ideal Use Case

An AI lab needing high-accuracy coding or agentic post-training data contracts Deccan to design tasks, build RL environments, and have domain experts produce and verify the datasets, then runs continuous evaluations through Deccan's Helix suite.

How Deccan AI differentiates

Against Scale AI, Deccan concentrates on harder, higher-skill post-training work (coding, math, agentic, robotics) rather than broad high-volume labeling, and sources experts primarily from India to manage quality and cost. Against Surge AI, which is known for premium human-feedback data, Deccan bundles RL environments and an evaluation suite alongside raw datasets. Against Mercor, which centers on matching expert talent to AI labs, Deccan delivers managed datasets and tooling rather than staffing alone. As a 2023-founded company with a smaller public footprint, it has less brand recognition and a shorter track record than the larger incumbents.

FAQ

Q: What does Deccan AI do? A: Deccan AI builds expert training data for AI models, including SFT and RLHF datasets, reinforcement-learning environments, model evaluations, and agentic solutions, with a focus on coding, math, agentic, and robotics tasks.

Q: Who founded Deccan AI? A: Deccan AI was founded in 2023 by Rukesh Reddy and is headquartered in the San Francisco Bay Area with a large operations team in Hyderabad, India.

Q: How much has Deccan AI raised? A: Deccan AI raised a $25 million Series A in March 2026, led by A91 Partners with participation from Susquehanna International Group and existing investor Prosus Ventures.

Q: Deccan AI vs Scale AI: what is the difference? A: Scale AI is a large, broad data-labeling provider, while Deccan AI focuses on higher-skill post-training data (coding, math, agentic, robotics) and bundles RL environments and evaluations, sourcing experts mainly from India.

Q: Who are Deccan AI's customers? A: Deccan AI has named Google and Snowflake among its customers and works with frontier labs and enterprises on post-training data and evaluation projects.

tl;dr

Deccan AI is an enterprise data services company providing expert SFT and RLHF data, RL environments, evaluations, and agents for training and post-training AI models. Founded in 2023 and backed by a $25M Series A from A91 Partners, SIG, and Prosus, it counts Google and Snowflake as customers. Best for AI labs needing accurate, high-skill post-training data.

Related

Looking for more options? Browse the AI/ML Models directory or read our best AI models listicle. Deccan AI is also tracked on Crunchbase.