Grafton Sciences
Senior RL Research Scientist
Grafton Sciences, Redwood City, California, United States, 94061
About Grafton Sciences
About Grafton Sciences
We’re building AI systems with general physical ability — the capacity to experiment, engineer, or manufacture anything. We believe achieving this is a key step towards building superintelligence. With deep technical roots and real-world progress at scale (e.g., a $42M NIH project), we’re pushing the frontier of physical AI. Joining us means inventing from first principles, owning real systems end-to-end, and helping build a capability the world has never had before. About The Role
About The Role
We’re seeking a Senior RL Research Scientist to design and train reinforcement learning systems that optimize tool control, process tuning, and long-horizon workflows. You’ll build RL environments grounded in real physics, simulation, and digital twins; integrate dense verifiers; design safe RL strategies; and drive the transition from offline data to robust online behavior. This role spans algorithm development, systems integration, and hands-on experimentation in complex, high-dimensional domains. Responsibilities
Build RL environments for optimization, process tuning, and tool orchestration using real-world simulation and digital twins. Design and implement safe RL methods, verifier-integrated rewards, offline→online transitions, and policy evaluation pipelines. Develop state representations, action abstractions, and constraint mechanisms for reliable long-horizon decision-making. Collaborate with LLM researchers, agent systems, simulation teams, and tooling engineers to deploy RL agents into real workflows. Qualifications
Strong background in reinforcement learning, optimal control, or sequential decision-making, with experience applying RL to complex real or simulated systems. Familiarity with safe RL, constrained RL, verifier/detector integration, or multi-step policy evaluation frameworks. Demonstrated ability to build RL environments, design reward structures, and diagnose policy behavior at scale. Comfortable working across ML, simulation, systems engineering, and physical-toolchain interfaces in a fast-paced research environment. Above all, we look for candidates who can demonstrate world-class excellence.
Compensation We offer competitive salary, meaningful equity, and benefits.
#J-18808-Ljbffr
About Grafton Sciences
We’re building AI systems with general physical ability — the capacity to experiment, engineer, or manufacture anything. We believe achieving this is a key step towards building superintelligence. With deep technical roots and real-world progress at scale (e.g., a $42M NIH project), we’re pushing the frontier of physical AI. Joining us means inventing from first principles, owning real systems end-to-end, and helping build a capability the world has never had before. About The Role
About The Role
We’re seeking a Senior RL Research Scientist to design and train reinforcement learning systems that optimize tool control, process tuning, and long-horizon workflows. You’ll build RL environments grounded in real physics, simulation, and digital twins; integrate dense verifiers; design safe RL strategies; and drive the transition from offline data to robust online behavior. This role spans algorithm development, systems integration, and hands-on experimentation in complex, high-dimensional domains. Responsibilities
Build RL environments for optimization, process tuning, and tool orchestration using real-world simulation and digital twins. Design and implement safe RL methods, verifier-integrated rewards, offline→online transitions, and policy evaluation pipelines. Develop state representations, action abstractions, and constraint mechanisms for reliable long-horizon decision-making. Collaborate with LLM researchers, agent systems, simulation teams, and tooling engineers to deploy RL agents into real workflows. Qualifications
Strong background in reinforcement learning, optimal control, or sequential decision-making, with experience applying RL to complex real or simulated systems. Familiarity with safe RL, constrained RL, verifier/detector integration, or multi-step policy evaluation frameworks. Demonstrated ability to build RL environments, design reward structures, and diagnose policy behavior at scale. Comfortable working across ML, simulation, systems engineering, and physical-toolchain interfaces in a fast-paced research environment. Above all, we look for candidates who can demonstrate world-class excellence.
Compensation We offer competitive salary, meaningful equity, and benefits.
#J-18808-Ljbffr