Trajectory, a startup founded by three Stanford alumni from Google DeepMind and Apple, has raised $45 million in Series A funding at a $300 million valuation. The company addresses the critical limitation of current AI models by automating the process of learning from feedback. Founder Ronak Malde compares foundational models to smart PhD students starting new jobs, arguing that the economy values experienced professionals.
Trajectory's platform enables businesses to connect AI models to their applications, collecting feedback to retrain and improve performance at specialized tasks over time. This approach, known as continual learning, allows nontechnical specialists like lawyers and doctors to improve models without hiring expensive research teams. Early customers include Harvey, Rogo, Decagon, Mercor, and Nvidia.
The startup aims to make specialized AI deployment faster and safer by keeping humans involved in the training process, differentiating itself from competitors like Thinking Machines Lab and Applied Compute. The 18-person team has recruited talent from OpenAI, Stripe, and Netflix. Trajectory recently made its post-training platform publicly available, allowing specialists to define what "good looks like" for their organizations.
Source: TechCrunch Venture · Summarized by HeadlinesBriefing