A bunch of AI researchers who beforehand labored at Google DeepMind, Apple, OpenAI, and Meta Superintelligence Labs introduced on Wednesday they’re launching a brand new startup known as Trajectory, which goals to assist firms recurrently enhance their AI merchandise by coaching on real-world consumer interactions.
Trajectory desires to construct a platform for AI that may be taught repeatedly, a functionality that researchers have lengthy held up as a significant barrier to additional AI progress. OpenAI, Google, and Anthropic have discovered success coaching more and more succesful variations of AI fashions, particularly for domains corresponding to coding, math, and science. Nonetheless, these techniques cease getting smarter after their coaching is completed. Whereas there have been some current breakthroughs in continuous studying, tech firms have usually struggled to make AI merchandise that be taught from their errors in actual time. In December 2025 at NeurIPS, one of many largest annual AI analysis conferences, Turing award winner Richard Sutton argued that continual learning is essential for constructing superintelligent brokers.
Trajectory has raised a $15 million seed spherical at a $115 million post-money valuation, led by the enterprise capital agency Conviction, with participation from Bessemer Enterprise Companions, Radical VC, and BoxGroup. Particular person traders additionally participated within the spherical, together with Google DeepMind’s chief scientist, Jeff Dean, in addition to the so-called “godmother of AI,” Stanford professor and World Labs CEO Fei-Fei Li.
Trajectory’s CEO and cofounder Ronak Malde was beforehand an AI researcher at Windsurf, and he later turned one among solely a handful of workers who went to work at Google DeepMind when it employed the coding startup’s high expertise in a $2.4 billion deal final yr. The opposite cofounders of Trajectory embrace Arjun Karanam, a former AI researcher at Apple who labored on the Imaginative and prescient Professional, and Michael Elabd, who beforehand labored in Google DeepMind’s robotics division.
Malde tells WIRED that some main AI coding merchandise, corresponding to Cursor, are already doing an early model of continuous studying—utilizing real data about how people interact with their merchandise to do post-training and recurrently ship mannequin enhancements. He argues this can be a core cause why AI coding merchandise have taken off so quickly, and is a part of the explanation why main AI labs have rushed to develop vibe coding functions of their very own. With Trajectory, Malde and his group of 11 researchers and engineers hope to use an analogous approach for bettering AI-powered instruments outdoors the coding area.
“Even essentially the most highly effective AI right now remains to be static. The AI mannequin that you simply used yesterday goes to make the identical errors right now,” says Malde. “A pair firms are beginning to get to that world of continuous studying. What we’re doing is constructing the platform for each single firm to get to continuous studying.”
The problem with making use of this logic to different domains is that coding is definitely verifiable—code both runs or it doesn’t—however some industries have looser definitions of success. Karanam says a part of what Trajectory’s platform gives helps optimize an AI mannequin to a enterprise’s particular wants.
Reasonably than ranging from an off-the-shelf mannequin from OpenAI or Anthropic, Trajectory has prospects start with an open-source mannequin that has been post-trained for a particular AI product the corporate has in thoughts. For Decagon, a buyer that builds AI buyer help brokers, Trajectory logs when its AI falls quick—say, a buyer attempting to make a return will get their question bounced to a human—and makes use of these situations to post-train a brand new mannequin as typically as each week. Trajectory claims these post-trained fashions beat the frontier labs’ fashions on slender duties that matter most for an organization’s product.
