Trajectory raises $45 million to make AI model training accessible
Trajectory, an artificial intelligence startup founded about a year ago by three Stanford classmates, has raised $45 million in a Series A funding round led by Sequoia, bringing the 18-person company's valuation to $300 million. Other participants in the round include Nvidia, Bessemer Venture Partners, and Bain Capital Ventures. Launched by Ronak Malde, Michael Elabd, and Arjun Karanam—who bring engineering and research backgrounds from Google DeepMind and Apple—the startup focuses on automating post-training so AI models can continuously learn and adapt to specialized tasks.
Adapting general AI models for complex industries like law and finance typically requires costly internal research teams or forward-deployed engineers. Trajectory addresses this limitation by offering an accessible continual-learning platform where businesses connect open-source models to their applications and retrain them using real-world performance feedback. Co-founder Michael Elabd stresses that keeping human experts in the training loop prevents the safety risks tied to fully autonomous recursive self-improvement, allowing professionals such as lawyers or doctors to directly shape how models improve.
Now opening its platform publicly, Trajectory already counts enterprise customers such as Harvey, Rogo, Decagon, Mercor, and Nvidia. The company enters an increasingly contested market for post-training automation, competing against Prime Intellect, the $3.25 billion startup Applied Compute, and Thinking Machines Lab, established by former OpenAI CTO Mira Murati.
Continue in the app — vote & join in ➔