Trajectory
Use application feedback to improve a model after its initial training.
Trajectory Funding History
Trajectory operates a continual-learning post-training platform that retrains AI models from application feedback.
Company Overview
Trajectory
Founded 2025
Official WebsiteTrajectory Funding Timeline
Series A
US$45 million
Trajectory builds tools that collect feedback from deployed applications and use it to retrain the models behind them.
Led by Sequoia Capital. Participation included NVIDIA, Bessemer, Bain Capital Ventures.
Source · techcrunch.com ↗
Trajectory Founding Team
Trajectory brought together researchers with experience at Google DeepMind and Apple. Their company focuses on post-training that uses feedback from deployed applications. The founding team included Ronak Malde, Michael Elabd, and Arjun Karanam.
Trajectory Products
Trajectory Company Development
Application Feedback as Training Material
Trajectory connects post-training to the feedback an application collects in use. This creates a route for teams to improve how a model handles a task after deployment, when actual user interactions can reveal gaps that were hard to anticipate before launch.
Trajectory Competitive Landscape
From Observing a Failure to Training on It
Raindrop helps an agent team find and investigate production failures. Trajectory starts from feedback that can be used to retrain the model. The distinction is whether the team needs to identify the issue first or has examples ready to feed into a new training cycle.
