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

Trajectory has raised $45 million in a Series A round at a $300 million valuation to automate the post-training of AI models. Sequoia led the round, with Nvidia, Bessemer Venture Partners, and Bain Capital Ventures participating.

The 18-person startup, founded about a year ago by three Stanford classmates, wants to fix a core limitation of today's AI models: they don't learn from their mistakes without further training. Businesses in fields like law and finance typically fine-tune general models from OpenAI, Anthropic, or open-source alternatives, but that usually means hiring internal research teams or paying expensive consultants known as forward-deployed engineers.

What does Trajectory's platform actually do?

Trajectory's approach is called continual learning. A business starts with an open-source AI model, connects Trajectory's platform to its applications, and lets the system collect feedback on how the model performs. Trajectory then uses that feedback to retrain the model, improving its performance at specialized tasks over time.

Co-founder Ronak Malde compared foundational models to a very smart PhD student who's always starting a new job. "But what the economy values is someone who worked on the job for 30 years," he told TechCrunch. With Trajectory, he said, lawyers, doctors, and other specialists can improve the models "near automatically" themselves, which could make deploying specialized AI faster and cheaper.

Customers decide "what good looks like for their organization and can turn that into improvements for their model," Malde said.

The startup is now making its post-training platform publicly available.

Who is already using it?

Harvey and Rogo are using Trajectory to post-train open-source models so they perform more like experienced lawyers or financial analysts. The models learn from previous tasks, such as reviewing a particular type of legal contract or valuing a company using a specific methodology. Trajectory's other customers include Decagon, Mercor, and Nvidia.

How is this different from recursive self-improvement?

Trajectory's continual learning approach shares similarities with recursive self-improvement (RSI), the idea that AI systems can improve their own capabilities. Co-founder Michael Elabd argues that keeping humans in the training loop makes Trajectory safer, because autonomous self-improvement could lead to unintended or even dangerous outcomes.

Trajectory isn't alone in the market. It counts Thinking Machines Lab — founded by former OpenAI CTO Mira Murati and the company behind the fine-tuning platform Tinker — as a direct competitor. Applied Compute, valued at $3.25 billion last month, and Prime Intellect, which helps companies build and train their own AI agents, are also competing in the post-training space.

Why does the team matter?

Malde believes the startup's biggest advantage is its people. He was among the handful of researchers Google DeepMind acqui-hired from Windsurf. Elabd also worked at DeepMind. The third co-founder, Arjun Karanam, previously worked on generative foundation models at Apple. The startup has also recruited researchers and product designers from OpenAI, Stripe, and Netflix.

Design matters to an AI training company for a practical reason, Karanam said: Trajectory has built an interface that makes post-training accessible to its non-technical customers. That interface is the bridge between specialists who know what "good" looks like and the models that need to learn it — and with $45 million in fresh capital, Trajectory now has the runway to push it into the open market.