Humanoid Whole Body Control Founding Engineer
Minerva Humanoids
San Francisco, CA, USA
USD 200k-300k / year + Equity
Posted on Apr 30, 2026
About the RoleMinerva Humanoids is developing rugged humanoid robots to do the most dangerous jobs on the planet. Our Whole-Body Controls group sits at the intersection of reinforcement learning, optimal control, and physical robotics. We train policies that learn a broad repertoire of whole-body skills and deploy them on real hardware in unstructured, high-consequence environments.As Founding Engineer on the Whole-Body Controls team, you will lead the design and investigation of learning-based control methods for life-saving applications. This is a research role with a strong deployment mandate: the ideas you develop will be tested on physical robots, not just in papers. You will have the freedom to define your own research agenda within the team’s mission, publish your work at top venues, and shape the technical direction of a core capability.What You’ll Work On- Formulate and investigate novel approaches to whole-body policy learning, including hierarchical, multi-task, and compositional architectures for skill orchestration across locomotion, manipulation, and transitional behaviors.- Develop methods to improve sim-to-real transfer for contact-rich, whole-body tasks, including system identification, domain randomization strategies, and techniques for adapting to hardware degradation and field wear.- Design rigorous experimental protocols to benchmark policy performance across diverse scenarios, terrains, and operator-directed tasks, and build the evaluation infrastructure to support rapid iteration.- Collaborate closely with the perception and hardware teams to co-design observation and action spaces, sensor configurations, and onboard compute pipelines that serve learned policies.- Publish and present your research at leading venues (CoRL, RSS, ICRA, NeurIPS, ICLR) and contribute to Minerva’s presence in the research community.- Mentor junior researchers and engineers from our academic partners, and help build a research culture grounded in scientific rigor and engineering pragmatism.What You’ll Bring- PhD in robotics, computer science, mechanical engineering, or a closely related field, with a dissertation focus on learning-based control, reinforcement learning for robotics, or whole-body motion planning.- A strong publication record at venues such as CoRL, RSS, ICRA, NeurIPS, ICLR, or equivalent, demonstrating original contributions to robot learning or control.- Hands-on experience deploying learned control policies on real legged robots (incl. quadrupeds or humanoids), with a deep appreciation for the gap between simulation results and physical performance.- Expertise in modern RL algorithms (PPO, SAC, model-based methods) and practical fluency with at least one major RL/simulation framework (IsaacLab, MJX, RSL-RL, or similar).- Strong software skills in Python and PyTorch. Comfort using LLM-based coding tools to accelerate iteration, and the judgment to critically evaluate their outputs.- Exceptional independence and scientific taste: you identify the right problems to work on, design clean experiments, and draw honest conclusions from the results.Nice to Have- Experience with optimization-based control methods (MPC, QP solvers, trajectory optimization) and a perspective on how they complement and constrain learned policies.- Prior work on contact-rich manipulation, dexterous grasping, or loco-manipulation on physical hardware.- Experience leading a small research team or co-advising graduate students.- Familiarity with hardware-in-the-loop testing, real-time control systems, or embedded deployment of neural network policies.Expected Compensation:$200,000 – $300,000 annual salary + 1–2% equity + benefitsPay offered may vary depending on multiple individualized factors, including market location, job-related knowledge, skills, and experience. The total compensation package for this position may also include other elements dependent on the position offered. Details of participation in these benefit plans will be provided if a candidate receives an offer of employment.