Trajectory secures $45M for automated AI model fine-tuning

Trajectory, a one-year-old startup founded by alumni from Google DeepMind and Apple, has raised $45 million to automate the continuous improvement of AI models. The Series A round was led by Sequoia, with additional investments from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures.

Currently, adapting foundational AI models for specialized industry applications—such as legal or financial tasks—requires significant technical resources or expensive engineering consultants. Trajectory's platform aims to bypass this by integrating with a company's existing applications to collect performance feedback and automatically retrain models. This approach focuses on 'continual learning,' where the AI improves its performance through repetitive interaction with real-world data.

Co-founder Ronak Malde noted that while current foundational models possess general intelligence, they lack the domain-specific experience gained over time. By enabling specialists to refine models without needing deep technical expertise, the startup intends to lower the barriers to deploying tailored AI solutions.

With an 18-person team, Trajectory plans to use the new capital to scale its operations. The company's goal is to make the deployment of specialized, high-accuracy AI models both faster and more cost-effective for businesses.