- Trajectory, an 18-person startup, raised a $45 million Series A at a $300 million valuation, led by Sequoia with Nvidia, Bessemer, and Bain Capital Ventures participating.
- The company’s continual-learning platform lets businesses retrain open source AI models using real feedback on specialized tasks, without hiring dedicated ML teams.
- Customers Harvey and Rogo already use Trajectory to post-train models to act more like experienced lawyers and financial analysts.
- Co-founder Michael Elabd says keeping humans in the training loop, unlike full autonomous self-improvement, keeps the approach safer.
What Happened
Trajectory, a startup founded about a year ago by three Stanford classmates with backgrounds at Google DeepMind and Apple, has raised a $45 million Series A at a $300 million valuation, led by Sequoia with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures (TechCrunch, Nov. 2, 2026). The 18-person company is also making its post-training platform publicly available for the first time.
Why It Matters
The round adds to a fast pace of AI funding MegaOne AI tracked in its roundup of 2026’s biggest rounds. Co-founder Ronak Malde compared general-purpose foundation models to “a very smart PhD student who’s always starting a new job,” arguing specialized businesses need models that retain experience the way a long-tenured employee would.
Technical Details
Trajectory connects to an open source model already deployed in a business’s applications, collects feedback on how it performs, and retrains it to improve at specialized tasks over time, an approach the company calls continual learning. Customers Harvey and Rogo use the platform to post-train open source models so they perform more like experienced lawyers or financial analysts on tasks such as contract review. Decagon, Mercor, and Nvidia are also customers.
Who’s Affected
Businesses in specialized fields like law and finance gain an alternative to hiring internal research teams or forward-deployed engineers to fine-tune models. Open source model providers benefit from a path to deeper enterprise customization without building proprietary retraining infrastructure.
What’s Next
Trajectory frames its approach as a safer alternative to recursive self-improvement, the idea that AI systems retrain themselves without human oversight. Co-founder Michael Elabd said keeping humans in the training loop avoids the unintended outcomes he associates with fully autonomous self-improvement.
