Trajectory launches self-serve platform for post-training models on production data
Ronak Malde's startup opened its self-serve platform on November 2nd with a $45M Sequoia-led Series A at a $300M valuation.
By Ryan Merket · Published
Primary source: x.com
Why it matters
Post-training is where general models become useful for a specific company's work, and it still usually requires scarce research talent. Trajectory is betting that expert judgment, captured through Continual Grading and trained on single production rollouts with GS1, can replace much of that labor - a claim Harvey, Rogo, Mercor, and NVIDIA are already testing, and one that puts the startup directly against Thinking Machines Lab's Tinker and Applied Compute.
Trajectory cofounder and CEO Ronak Malde (@rronak_) launched the self-serve Trajectory platform for post-training on November 2nd, alongside a $45M Series A led by Sequoia at a $300M valuation. The round included NVIDIA, Bessemer, and Bain Capital Ventures.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.