Join a frontier AI team building systems that can act in the physical world, experimenting, optimizing, and controlling real processes through advanced ML, simulation, and automation. This group is pushing the boundaries of physical intelligence, backed by significant long-term funding and a mandate to invent from first principles.
If you want to:
- Work on problems few teams in the world can touch
- Build RL systems that power real tools, workflows, and scientific processes
- Operate in a fast, high-ownership, deeply technical culture
The Role
You’ll design and deploy reinforcement learning systems that control complex tools, optimize multi-step processes, and operate across high-fidelity simulations and digital twins. Expect hands-on research, real-world experimentation, and tight collaboration with teams across ML, simulation, and systems engineering.
What You’ll Do
- Build RL environments for tool control, workflow optimization, and long-horizon decision-making
- Develop safe and constrained RL methods, verifier-driven rewards, and offline to online training pipelines
- Create state/action representations and evaluation frameworks for reliable policy behavior
- Work with cross-functional researchers and engineers to deploy RL agents into real workflows
What You Bring
- Strong background in RL, optimal control, or sequential decision-making
- Experience applying RL to complex simulated or physical systems
- Familiarity with safe/constrained RL, verifiers, or advanced evaluation pipelines
- Ability to design environments, rewards, and diagnostics at scale
- Comfort working across ML, simulation, and systems interfaces
