LLM Researcher at well-funded seed-stage robotics AI startup

You will lead the development of foundation models for robotics, moving beyond hand-engineered controllers to create general-purpose autonomous policies. By applying advanced post-training techniques to video diffusion models, you will help build systems that generalize across tasks and environments. This role is ideal for hands-on researchers who ship production-grade code.

Well-funded seed-stage robotics AI startup

• Execute end-to-end post-training workflows including SFT, RLHF, and DPO to align robot policy behaviors.
• Develop and optimize scalable training pipelines for large-scale video diffusion and foundation models.
• Rapidly iterate on model architectures and training infrastructure to improve zero-shot generalization in new environments.

• Extensive hands-on experience in post-training LLMs, specifically focused on alignment, reasoning, or multi-modal learning.
• Prolific coding ability with a track record of shipping research code, open-source contributions, or production ML systems.
• Hungry, early-career researcher excited by the prospect of working in-person in San Francisco within a fast-moving, high-stakes startup.

• Work at the cutting edge of physical AI by applying LLM-style scaling and generalization to humanoid robot autonomy.
• Backed by world-class AI researchers and top-tier venture capital firms at the earliest, highest-impact stage of growth.
• High-intensity, high-ownership environment where your research directly dictates the capabilities of next-generation robotic hardware.