Source : TechCrunch
SAN FRANCISCO, November 2, 2026 – Trajectory, a year-old artificial intelligence start-up, has raised US$45 million in a Series A funding round that values the company at US$300 million, as reported by TechCrunch and shared on X.
The round was led by Sequoia, with Nvidia, Bessemer Venture Partners and Bain Capital Ventures also taking part. Alongside the funding, the 18-person company is making its post-training platform publicly available.
Trajectory is tackling a well-known weakness of today’s AI models, whether they come from large labs or open-source projects: they do not learn from their mistakes unless they are trained further. Businesses that adapt general models for specialised fields such as law and finance usually fine-tune them, which often means hiring in-house research teams or costly outside consultants known as forward-deployed engineers.
The start-up wants to automate much of that work through an approach called continual learning. A business starts with an open-source model and connects Trajectory’s platform to its own applications. The platform gathers feedback on how the model performs in real use and then uses that feedback to retrain it, so the model gets better at specialised tasks over time.
Co-founder Ronak Malde told TechCrunch that foundation models are like a very smart PhD student who is always starting a new job, whereas the economy values someone who has been in the job for 30 years. He said lawyers, doctors and other specialists using Trajectory can improve models almost automatically themselves, deciding “what good looks like for their organization” and turning that into improvements, which could make deploying specialised AI faster and cheaper.
Legal AI company Harvey and financial research firm Rogo already use the platform to post-train open-source models so that they perform more like experienced lawyers or financial analysts, learning from earlier tasks such as reviewing a certain type of contract or valuing a company with a particular method. Other customers include Decagon, Mercor and Nvidia.
Trajectory’s approach has some similarities with recursive self-improvement, the idea that AI systems can raise their own capabilities. Co-founder Michael Elabd argued that keeping humans involved in the training process makes the company’s approach safer, since fully autonomous self-improvement could lead to unintended or even dangerous outcomes. The company’s website says every model update goes through a customer’s own evaluation and approval process before it reaches production.
The firm was founded by three Stanford classmates. Malde was among the researchers Google DeepMind hired from coding start-up Windsurf, Elabd also worked at DeepMind, and third co-founder Arjun Karanam previously worked on generative foundation models at Apple. Karanam said the company has built an interface designed to make post-training easier for customers without technical backgrounds.
The new round builds on a US$15 million seed round announced in May 2026, when Trajectory introduced itself as a research lab and product company focused on continual learning. Its website says the platform plugs into a product through a lightweight software development kit and captures signals that users already generate, such as traces, corrections, re-prompts and edits, while customers keep full control over which data is used for training.
Trajectory faces competition in the race to automate post-training, including from Thinking Machines Lab, the company founded by former OpenAI chief technology officer Mira Murati, as well as Applied Compute and Prime Intellect.
About The Author
