Trajectory, an 18-person startup that automates AI model post-training, has raised a $45 million Series A at a $300 million valuation led by Sequoia — and made its platform publicly available. Why it matters: the round is a bet that the next bottleneck in enterprise AI is not building models, but teaching the ones companies already have to get better at their specific jobs without hiring a research team.
November 2, 2026
What happened
Trajectory announced on Monday that it has raised $45 million in a Series A round valuing the company at $300 million. Sequoia led the round, with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures. The company was founded about a year ago by three Stanford classmates — Ronak Malde, Michael Elabd, and Arjun Karanam — and now employs 18 people. Alongside the funding, Trajectory is making its post-training platform publicly available for the first time.
How the platform works
Trajectory’s pitch starts from a limitation every current model shares: once a model ships, it does not learn from its mistakes unless someone trains it again. Companies in fields like law and finance usually fix that by fine-tuning, which often means hiring internal researchers or expensive forward-deployed engineers. Trajectory automates that loop through what it calls continual learning. A business starts with an open source model, connects Trajectory’s platform to its applications, and the platform collects feedback on how the model performs in real use. That feedback is then used to retrain the model so it improves at the organization’s specialized tasks over time.
Co-founder Ronak Malde compared foundation models to a very smart PhD student who is always starting a new job. “But what the economy values is someone who worked on the job for 30 years,” he told TechCrunch. The idea is that the specialists themselves — lawyers, doctors, analysts — define what good work looks like and turn that judgment into model improvements, near automatically, rather than waiting on an engineering team.
Who is already using it
Harvey and Rogo are already using the platform to post-train open source models, teaching them to work more like experienced lawyers or financial analysts — for example, reviewing a particular type of legal contract or valuing a company using a specific methodology. Trajectory’s other named customers include Decagon, Mercor, and Nvidia. The founding team is part of the draw: Malde was among the researchers Google DeepMind acqui-hired from Windsurf, Elabd also worked at DeepMind, and Karanam previously worked on generative foundation models at Apple. The company has also recruited from OpenAI, Stripe, and Netflix.
Trajectory is not alone in this market. It names Thinking Machines Lab — the company founded by former OpenAI CTO Mira Murati, behind the fine-tuning platform Tinker — as a direct competitor, along with Applied Compute, which was valued at $3.25 billion last month, and Prime Intellect. Elabd also draws a line between Trajectory’s approach and recursive self-improvement: keeping humans involved in the training process, he argues, makes the system safer than models improving themselves autonomously.
Why it matters
For buyers, this is the “picks and shovels” layer of AI getting serious money. Most companies will never train a frontier model, but many would pay to make a cheaper open model behave like a veteran employee in their niche — if they can do it without hiring the scarce engineers who know how. A $300 million valuation for an 18-person company says investors believe post-training automation is a real category, not a feature. The test now is whether continual learning from everyday feedback produces gains customers can measure, rather than models that quietly drift. If it works, the advantage in enterprise AI shifts a little further from whoever has the biggest model to whoever learns fastest from their own work.
Frequently asked questions
What does Trajectory do?
Trajectory sells a platform that automates post-training for AI models. Businesses connect it to an open source model running in their applications; the platform collects feedback from real use and retrains the model so it improves at specialized tasks over time — a process the company calls continual learning.
How much did Trajectory raise, and who invested?
Trajectory raised $45 million in a Series A round at a $300 million valuation. Sequoia led the round, with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures.
Who uses Trajectory’s platform?
Named customers include Harvey, Rogo, Decagon, Mercor, and Nvidia. Harvey and Rogo use it to post-train open source models for legal and financial work. The platform became publicly available alongside the funding announcement.
How is Trajectory different from fine-tuning services like Tinker?
Trajectory positions its platform around continual learning from live feedback, designed so nontechnical specialists — lawyers, doctors, analysts — can drive improvements themselves. It names Thinking Machines Lab’s Tinker, Applied Compute, and Prime Intellect as competitors in automated post-training.
Sources: TechCrunch