Trajectory Raises $45M Series A at a $300M Valuation

Trajectory raised $45 million in a Series A funding round at a $300 million valuation, according to FinSMEs' March 11, 2026 report on the San Francisco startup. Sequoia Capital led the investment. Nvidia, Bessemer Venture Partners and Bain Capital Ventures also participated. The round brought Trajectory's total funding to $60 million.

What does Trajectory build?

FinSMEs described Trajectory as a San Francisco, CA-based developer of continuous post-training and alignment software for artificial intelligence models, founded by Ronak Malde in 2026. FinSMEs said the company's specialized machine learning platform lets companies post-train frontier AI models for domain-specific tasks.

The platform also creates a closed-loop system, FinSMEs reported. Deployed models collect real-world production feedback. The models then improve autonomously from that data. The system makes live use part of the development cycle rather than a final check after deployment.

How will the money be used?

FinSMEs said Trajectory will direct the $45 million toward four priorities:

• Scale infrastructure for continual machine learning.
• Fund ongoing research and development for its optimization algorithms.
• Expand its deep-tech engineering workforce.
• Accelerate the commercial rollout of its enterprise model alignment platform.

The spending plan connects infrastructure, research, hiring and commercialization to the same objective: moving the platform from research previews into wider enterprise use. It also gives Trajectory room to expand while it develops the algorithms that underpin continual learning.

How does the learning loop work?

Trajectory's pitch starts with adapting a frontier model to a company's own work, FinSMEs said. The system then uses feedback from live production to support further improvement, creating an operating cycle rather than a training project that ends at launch. It resembles a continuous quality-control line: each deployment supplies evidence for another model iteration.

That loop depends on measurement as much as training. Continual Grading, one of two foundational research previews, is a framework designed to scale automated evaluation metrics, FinSMEs said. The preview supports assessment as production feedback accumulates.

Its second preview targets model scale and live learning. GS1, a proprietary algorithm, allows trillion-parameter models to learn directly from active deployment data, according to FinSMEs. In practical terms, the algorithm lets a deployed frontier model learn from activity it encounters after release.

Where could the technology be used?

FinSMEs said the two previews allow industries including law, medicine and engineering to build custom, proprietary value functions. The focus extends beyond generic scoring because companies can define outcomes relevant to their own domains, based on the source's description of domain-specific post-training.

Trajectory's approach offers a route beyond one-time model tuning. Instead of ending the training cycle before deployment, the system proposes a continuing process in which production feedback feeds model improvement. The commercial test is whether Trajectory can turn that architecture into a repeatable enterprise offering as its infrastructure and workforce expand.

What does the financing mean?\n

Based on figures reported by FinSMEs, the $45 million Series A represents three-quarters of Trajectory's $60 million in cumulative funding. It is also three times the $15 million the company had raised before this round. That concentration makes execution the immediate focus.

The $300 million valuation now sits against a company founded in 2026, according to FinSMEs. That short history makes the funding round a test of whether research previews can become a scaled commercial platform. The answer will depend on infrastructure, research, hiring and commercial execution, the four priorities supported by the proceeds.

What must Trajectory prove next?\n

The next test is not another model release alone. FinSMEs said Trajectory will use the capital to scale infrastructure, fund algorithm research, add deep-tech engineers and accelerate enterprise rollout. The financing shifts the burden from demonstrating the closed-loop design to showing that it can operate at commercial scale across enterprise deployments.