Trajectory Secures $45M to Automate AI Fine-Tuning

The company recently closed a $45 million Series A funding round, pushing its valuation to $300 million. Led by Sequoia, the investment also includes backing from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures. With an 18-person staff, the startup aims to simplify the transition from general-purpose AI to models that function like seasoned professionals in fields such as law and finance.

Trajectory’s platform utilizes a continual learning approach, enabling users to connect open-source models to their internal applications. As the system collects feedback, the platform retrains the model to handle specific tasks more effectively. Co-founder Ronak Malde notes that this allows lawyers and doctors to improve their tools without relying on specialized research teams. Early adopters, including Harvey, Rogo, and Decagon, are already using the software to refine performance on complex workflows like contract review and financial valuation.

While the platform shares conceptual ground with recursive self-improvement, the leadership team emphasizes a human-in-the-loop requirement to ensure safety. The startup now faces competition from well-funded rivals like Applied Compute and Thinking Machines Lab. To differentiate itself, the company is prioritizing an interface designed for non-technical users, attempting to lower the barrier for those who want their AI to reflect specific organizational standards.