Former Google and Apple researchers launch Trajectory, a startup building a platform for AI to learn continuously from real-world user interactions.

Trajectory raised a $15 million seed round at a $115 million post-money valuation, led by Conviction with participation from Bessemer Venture Partners, Radical VC, and BoxGroup. Individual investors included Google DeepMind chief scientist Jeff Dean and Stanford professor Fei-Fei Li.

The platform helps companies post-train AI models on real data about how users interact with their products—for example, Decagon's customer support AI logs when queries fail and retrains weekly. Unlike off-the-shelf models from OpenAI or Anthropic, Trajectory starts with open-source models customized for a company's specific task.

AI coding products like Cursor already practice this continual learning approach, which Trajectory's CEO argues is why they have grown rapidly. Trajectory aims to apply the same technique to other domains where success metrics are less clear-cut than in coding.