Trajectory Raises $45M Series A to Automate AI Model Training for Non-Technical Experts

Funding Details

Trajectory, a startup founded approximately one year ago, announced it has raised $45 million in a Series A funding round. The round was led by Sequoia, with additional participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures. This funding round values the company at $300 million.

Addressing AI Model Limitations

Current AI models, whether proprietary or open-source, require further training to learn from mistakes. Businesses typically fine-tune these models for specialized fields, which often necessitates hiring internal research teams or expensive forward-deployed engineers. Trajectory aims to automate this process, enabling AI models to continuously improve at specialized tasks based on feedback.

Continuous Learning Approach

Trajectory's platform implements a continual learning approach. Businesses can integrate an open-source AI model with the platform, which then collects feedback on the model's performance within their applications. This information is used to retrain the model, leading to improvements in specialized tasks over time. This process allows specialists like lawyers and doctors to improve models near-automatically.

Current Adoption and Availability

Companies such as Harvey and Rogo are already utilizing Trajectory's platform for post-training open-source models, helping them perform tasks like reviewing legal contracts or valuing companies. Other customers include Decagon, Mercor, and Nvidia. Trajectory is now making its post-training platform publicly available.