New AI lab Core Automation 'nerdsniped' researchers from Anthropic, Google DeepMind

Core Automation, a new AI startup founded by a former OpenAI researcher, has recruited senior talent from Anthropic and Google DeepMind. The company confirmed to Business Insider that it has hired Anil Kumar from Anthropic and one other researcher who led the CoRAG reasoning team at Google DeepMind. Kumar worked on Claude's core technical architecture, including what Anthropic describes as Claude's 'character' and safety guardrails. CoRAG is a technique that enables models to consult external knowledge bases during reasoning.

The move comes amid broader talent migration patterns in AI. Core Automation positions itself as aiming to develop agents that can 'complete complex work', with a founding vision focused on efficiency rather than generality. Founder Aditya Singh, previously a co-author on key OpenAI capabilities, stated that the lab will not aim to build a 'frontier GPT-class model', but instead focus on tailored, agentic solutions for enterprise use cases. The recruits were attracted, Singh said, by the promise of 'maximum focus, minimum friction' in the development process.

The announcements follow a pattern of high-profile researcher moves between leading AI labs, raising questions about knowledge continuity and the direction of model development. An Anthropic spokesperson declined to comment on the departures, citing their policy on internal personnel matters. Neither Kumar nor the unnamed DeepMind researcher have publicly commented on the career change. Core Automation has yet to release a product.