Trajectory has raised 45 million dollars in new funding that values the company at 300 million dollars. The startup, which is about a year old, is building software meant to let specialists without engineering backgrounds train artificial intelligence models. TechCrunch first reported the round.
The commercial idea is straightforward. People who understand a hospital ward, a warehouse, or a legal archive often know what a model should notice, yet they rarely control how that model is trained. Trajectory is selling a way for those experts to work with their own data without waiting on a machine learning group to write training pipelines.
That gap has become a bottleneck as companies try to adapt general systems to private records and internal rules. The work still tends to demand scarce engineers, costly computing time, and long review cycles. Money has kept flowing to firms that say they can shorten the path, even as buyers weigh accuracy, bias, and who is liable when a model fails.
Whether specialists who are not engineers can produce models that hold up once they leave a demonstration is the test ahead.
Little else about the round is public. The report that disclosed the financing did not name the investors or describe how Trajectory plans to spend the capital. The valuation itself is the clearest signal: 300 million dollars after roughly a year, a price that shows how much capital is still chasing tools built around how models are trained and tuned.
The harder question is whether specialists who are not engineers can produce models that hold up once they leave a demonstration. Domain knowledge does not by itself create sound evaluation, monitoring, or security. Trajectory now has more capital with which to try to close that gap, in a market already crowded with similar claims.
