Real stakes. Real-world impact.
The most interesting, and messiest, problems show up at the edges.
Gray Swan has made a habit of being first in line. We’re often the first to spot new failure modes, and, by extension, the first to defend against them. Being research-first isn’t a tagline. It’s how we build our products, our playbooks, and our culture of thinking three steps ahead.
Whether it’s stress-testing the latest agentic AI in a lab or helping enterprises deploy at full speed, our goal is the same: uncover the hard problems before anyone else does, then turn that insight into real-world solutions that let innovation happen, safely and without slowing down.
Our research has directly informed the safety evaluations of some of the most advanced AI models in the world.
As a Swan, individual work matters, immediately.
Your work helps shape how AI is deployed, evaluated, and trusted; from enterprises shipping to production, to researchers pushing new capabilities, to organizations responsible for safety at scale.
This is work with real consequences. Looking for a place where your thinking carries weight and where progress is measured in outcomes, not meetings? You’ll find it here.
Our benefits are designed to support real life, sustained curiosity, and the kind of deep work that meaningful impact requires.
If you’re excited by meaningful problems, real responsibility, and work that shows up in the world, you’ll fit right in. Take a look at our open roles below.
Join our Talent Newsletter and get notified about new roles.
