Job Description

Senior LLM Research Scientist

We’re hiring a Senior LLM Research Scientist to join a well-funded, frontier physical AI startup building systems with general physical ability — AI that can experiment, engineer, and manufacture in the real world.

The team is pushing LLMs far beyond chat interfaces and into autonomous, tool-using agents that drive real engineering workflows. These systems operate inside a 20,000 sq ft physical intelligence factory and support a $42M federally funded program , representing one of the most advanced real-world deployments of reasoning-first AI today.

This is a small, world-class algorithmic team working from first principles to build long-horizon reasoning, planning, and alignment frameworks for AI systems with real-world consequences.

What you’ll do:

• Build advanced planning, decomposition, and structured tool-use frameworks for high-agency LLM agents
• Lead SFT, RLHF/DPO, verifier-guided RL , and routing strategies to unlock major capability gains
• Design schemas, tool-call policies, safety constraints, and failure-recovery mechanisms for autonomous agents
• Drive integration with agent systems to evaluate LLM behavior in real engineering and manufacturing loops
• Work closely with researchers, systems engineers, and infrastructure teams in a fast-moving research environment

What we’re looking for:

• Senior or Principal-level experience in LLM research, reasoning models, agent systems, or structured tool use
• Strong background in SFT, RLHF, DPO, or verifier-driven RL , with demonstrated impact on model capabilities
• Deep understanding of long-horizon planning, policy constraints, and behavioral evaluation
• Comfortable operating across ML research, systems engineering, and real-world toolchains
• Thrives in high-ownership, first-principles, frontier research environments

Location: Redwood City, San Francisco (Onsite preferred)

Job Tags

Similar Jobs