Trajectory Secures $45M to Automate AI Training for Specialists
Startup Trajectory raises a $45 million Series A round to help nontechnical professionals continuously train and improve specialized AI models.
Modern artificial intelligence models developed by industry leaders often struggle with a fundamental drawback: they fail to learn from errors unless they undergo additional training. Organizations tailoring these general systems for niche sectors like law and finance usually resort to hiring dedicated research teams or expensive forward-deployed engineers.
Trajectory, a startup established roughly a year ago by three Stanford alumni with backgrounds at Apple and Google DeepMind, has emerged to automate this procedure. The 18-person enterprise recently secured $45 million in a Series A funding round, placing its valuation at $300 million. Sequoia led the investment, with backing from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures.
Solving the Continual Learning Challenge
Ronak Malde, co-founder of Trajectory, likened foundational models to brilliant recent graduates who lack practical experience. He noted that the market values long-term job experience instead. To bridge this gap, the startup utilizes continual learning. Companies integrate an open-source model with Trajectory’s system, which gathers operational feedback and retrains the model over time to handle specific workflows better.
This mechanism allows professionals such as doctors and lawyers to enhance AI systems automatically without deep technical expertise. Early adopters including Harvey, Rogo, Decagon, Mercor, and Nvidia already utilize the platform for post-training open-source models. For tasks like reviewing legal documents or executing financial valuations, organizations can define quality standards and feed those improvements directly into their models. While this shares conceptual traits with recursive self-improvement, co-founder Michael Elabd emphasizes that keeping humans in the loop maintains safety by avoiding uncontrolled autonomous changes.
Market Position and Team Expertise
Trajectory faces competition from firms like Applied Compute, Prime Intellect, and Thinking Machines Lab, the creator of the Tinker fine-tuning platform founded by former OpenAI CTO Mira Murati. However, Trajectory’s founders believe their talent pool provides a distinct edge. The founding team includes Malde and Elabd from DeepMind alongside Arjun Karanam, who previously developed generative foundation models at Apple. They have also added talent from OpenAI, Stripe, and Netflix.
Karanam highlights that strong design choices are central to their product, enabling a user-friendly interface that simplifies post-training for non-technical clients.
What it means for developers
As the ecosystem for post-training and customized models evolves, developers building specialized workflows need flexible, cost-effective infrastructure. Developers can try top AI models cheaply through one API at https://apixoai.online. As platforms like Trajectory make model refinement more accessible to domain experts, developers will likely see increased demand for seamlessly integrating feedback loops and tailored open-source solutions into business applications.
Source: Trajectory raises $45M to make AI model training accessible to nontechnical experts — TechCrunch AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
