Trajectory Raises $45M for AI Model Training Platform

Trajectory, an 18-person startup founded by three Stanford classmates with backgrounds at Google DeepMind and Apple, has raised $45 million in Series A funding to make AI model training accessible to domain experts without technical backgrounds. The round was led by Sequoia Capital, with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures, valuing the company at $300 million.

The Problem: AI Models That Can't Learn on the Job

Current AI models from OpenAI, Anthropic, and open-source alternatives share a fundamental limitation: they don't learn from their mistakes without additional training. Co-founder Ronak Malde compared foundational models to a brilliant PhD student always starting fresh. "But what the economy values is someone who worked on the job for 30 years," he explained.

Businesses deploying AI for specialized fields like law and finance typically solve this by fine-tuning general models, a process that requires hiring internal research teams or expensive forward-deployed engineers. Trajectory aims to automate this costly process through continual learning, allowing models to improve from feedback without constant human intervention.

How Trajectory's Platform Works

Trajectory enables businesses to start with an open-source AI model and connect the platform to their applications. The system collects feedback on model performance and uses that information to retrain the model automatically, helping it improve at specialized tasks over time.

Legal firms using Harvey and financial analysts working with Rogo are already leveraging Trajectory's platform to post-train open-source models. These models learn to perform more like experienced professionals by processing previous tasks, such as reviewing specific contract types or applying particular valuation methodologies. Other customers include Decagon, Mercor, and Nvidia.

Release Date and Public Availability

Officially launched on November 2, 2026, Trajectory is now making its post-training platform publicly available to enterprises. The platform allows specialists in law, medicine, and other fields to improve models themselves, determining "what good looks like for their organization and can turn that into improvements for their model," according to Malde.

Continual Learning vs. Recursive Self-Improvement

Trajectory's approach shares similarities with recursive self-improvement (RSI), where AI systems enhance their own capabilities. However, co-founder Michael Elabd argues that keeping humans involved in the training loop makes Trajectory's approach safer, as fully autonomous self-improvement could lead to unintended or dangerous outcomes.

The startup faces competition from Thinking Machines Lab (founded by former OpenAI CTO Mira Murati), Applied Compute (valued at $3.25 billion in October 2026), and Prime Intellect. Malde believes Trajectory's competitive advantage lies in its team: he was among researchers Google DeepMind acqui-hired from Windsurf, Elabd also worked at DeepMind, and third co-founder Arjun Karanam previously developed generative foundation models at Apple. The team includes researchers and designers from OpenAI, Stripe, and Netflix.

What This Means

Trajectory's funding and public launch signal growing demand for automated AI training tools that reduce deployment costs and timelines. By enabling nontechnical domain experts to improve specialized models through continual learning, the platform could accelerate AI adoption in industries where accuracy and domain-specific knowledge are critical. The $45 million investment validates the market opportunity for post-training automation as enterprises move beyond general-purpose models toward specialized, continuously improving AI systems.