Patronus AI, a startup founded in 2023 by former Meta AI researchers Anand Kannappan and Rebecca Qian, announced a $50 million Series B funding round. The round, led by Greenfield Partners, includes participation from Notable Capital, Lightspeed, Datadog, and Samsung, bringing the company’s total funding to $70 million. The funding will support the development of digital environments to evaluate AI agents’ performance across complex tasks.
The company uses what it calls 'digital world models' to create replicas of websites and internal systems, allowing AI agents to be stress-tested after training using reinforcement learning. This method iteratively rewards successful task completion and penalizes errors, providing a more realistic test environment for AI agents. AI labs find these simulations valuable because they enable agents to encounter unpredictable scenarios, much like how Waymo trained autonomous vehicles using synthetic worlds to test against rare hazards.
Patronus is currently offering its simulated environments for software engineering and finance, but the company plans to expand into areas that are harder to verify. According to Kannappan, the focus remains on verifiable problems, but the long-term goal is to create environments where agents can operate for extended periods. The company also emphasizes that it competes primarily against the internal teams of AI labs rather than human-data firms like Mercor and Surge.
Source: techcrunch
