Trajectory Secures $45M Series A to Simplify AI Model Training for Specialists

AI models from major developers share a common shortcoming: they don't learn from errors without additional training. Businesses in fields like law and finance often fine-tune general models, but this typically means hiring internal research teams or costly consultants known as forward-deployed engineers.

Trajectory, a startup founded roughly a year ago by three Stanford classmates with backgrounds at Google DeepMind and Apple, aims to automate that process so models can learn from feedback and keep improving on specialized tasks.

The 18-person company announced a $45 million Series A round at a $300 million valuation, led by Sequoia with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures.

Co-founder Ronak Malde likened foundational models to a brilliant PhD student who keeps starting new jobs, while what the economy values is someone with 30 years on the job. Trajectory's approach, continual learning, lets businesses start with an open source model, connect the platform to their applications, and collect performance feedback that is used to retrain the model. Malde said lawyers, doctors, and other specialists can improve models nearly automatically themselves, potentially making specialized AI faster and cheaper to deploy.

Harvey and Rogo already use Trajectory's platform to post-train open source models, helping them perform more like experienced lawyers or financial analysts by learning from prior tasks such as reviewing a type of legal contract or valuing a company with a specific methodology. Other customers include Decagon, Mercor, and Nvidia. The startup is now making its post-training platform publicly available.

Malde said users decide what good looks like for their organization and can turn that into model improvements. Co-founder Michael Elabd argued that keeping humans involved makes the approach safer than autonomous self-improvement, which could produce unintended or dangerous outcomes.

Trajectory considers Thinking Machines Lab, Applied Compute, and Prime Intellect competitors. Malde sees the team as its biggest advantage: he was among researchers Google DeepMind acqui-hired from Windsurf, Elabd also worked at DeepMind, and third co-founder Arjun Karanam previously worked on generative foundation models at Apple. The startup has also recruited from OpenAI, Stripe, and Netflix. Karanam said the company built an interface making post-training more accessible to nontechnical customers.

Source: TechCrunch