Deccan AI has raised $25 million (about ₹235 crore) in a funding round led by A91 Partners, with participation from Susquehanna International Group and Prosus Ventures, underscoring investor faith in startups building core infrastructure for enterprise-grade artificial intelligence.
Funding underscores demand for post-training AI infrastructure
The capital injection reflects growing market demand for solutions that go beyond model development to address training data quality, evaluation and deployment readiness. As enterprises move from pilot projects to large-scale AI deployments, reliable systems for validating and optimising models have become essential.
Focus on post-training and evaluation
Founded in 2023 by Rukesh Reddy, Deccan AI specialises in improving model performance after initial training. Rather than building foundational models, the company develops tools and processes for refining outputs, boosting accuracy and ensuring models function robustly in real-world scenarios.
Its product suite includes reinforcement learning environments, evaluation platforms and workflow-based AI agents that enable teams to simulate operational conditions, surface errors and iterate faster. These capabilities are particularly relevant for regulated and high-stakes sectors such as finance, healthcare and industrial automation, where precision and safety are critical.
Early traction and international clientele
Deccan AI has secured engagements with global organisations, notably Google DeepMind and Snowflake, signalling demand from major technology firms for specialised training and evaluation services. The startup also leverages India’s engineering talent and domain expertise to create high-quality datasets, reinforcing the country’s role in the global AI value chain.
Plans for growth
With the new funding, Deccan AI plans to broaden its product roadmap, step up research efforts and scale its enterprise infrastructure. As AI adoption accelerates across industries, companies focused on post-training validation and reliability are expected to become integral to safe, scalable AI deployment.
