New startup from Stanford alumni raises $45 million

AI startup Trajectory has raised $45 million in a Sequoia-led Series A at a $300M valuation, opening access to its automated model fine-tuning platform designed for continuous learning from feedback.

Today's AI systems have a fundamental weakness: they can't learn from their mistakes automatically without additional training. When companies try to adapt these models to specialized domains, they are often forced to hire internal research teams or expensive external consultants.

How the Trajectory platform works

Trajectory wants to automate this process. Its platform allows businesses to gather information about an AI system’s performance from its use. The system then uses that information to retrain itself, improving its performance over time.

According to its creators, this process allows professionals to improve models almost automatically on their own, making the use of specialized artificial intelligence faster and more cost-effective. Companies such as Harvey, Rogo, Decagon, Mercor and Nvidia are already using Trajectory's platform, while the startup is now making its technology publicly available.

Security and competition

Co-founders Ronak Malde, Michael Elabd and Arjun Karanam note that keeping humans in the training process makes their approach safer, avoiding the unpredictable results of completely autonomous improvement. Trajectory competes with other companies in the space, including Thinking Machines Lab, Applied Compute and Prime Intellect.

↳

READ ALSO

HACKS.GR / RELATED

AI Unlocked

Apple's stock surges to $14,5 billion on memory prices

↗

>Stay connected with Hacks.gr

Set us as a preferred source on Google and find us on social media.

Preferred source

↗

end of article

Read the original English article on Hacks.gr