Trajectory has raised $45 million in a Series A funding round at a $300 million valuation to automate AI model post-training and make continual learning accessible to nontechnical experts. The 18-person startup, founded about a year ago by former Google DeepMind and Apple researchers, secured the funding in a round led by Sequoia, with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures.
Current foundational AI models often struggle to learn from their mistakes without expensive fine-tuning or internal research teams. Trajectory addresses this limitation by connecting open-source AI models to business applications, collecting performance feedback, and automatically retraining the models for specialized tasks in fields like law and finance. Companies including Harvey, Rogo, Decagon, Mercor, and Nvidia already use the platform to post-train models using human-in-the-loop oversight to ensure safety.
By providing an intuitive interface for specialists like doctors and lawyers, Trajectory aims to make specialized AI deployment faster and cheaper. Why it matters: This funding enables Trajectory to make its post-training platform publicly available, lowering the barrier for nontechnical businesses to continuously improve AI models.