Manipulation Capabilities Engineer

Responsibilities

• Post-train manipulation policies via behaviour cloning and RL
• Develop data preprocessing strategies to improve data quality
• Work with simulation teams to set up RL training using digital twins
• Partner with data collection teams to drive activities for specific capabilities
• Expand observation and action spaces for novel capabilities
• Partner with Teleoperations and Controls teams to improve motion smoothness
• Interface with hardware design teams to inform future hardware designs

Requirements

• 3+ years of experience working on robots in industry or research
• Experience with neural network post-training
• Familiarity with deep learning infrastructure (PyTorch, JAX, distributed training)
• Understanding of modern teleoperation and low-level control stacks
• Proficiency in modern software engineering practices
• Ability to document experiments and communicate trade-offs

Preferred Qualifications

• Experience training VLA models (autoregressive, diffusion, or flow-matching)
• Familiarity with OpenVLA or Physical Intelligence (π) models
• Experience applying RL to robotics problems
• Publications at top-tier robotics or deep learning conferences

Benefits

• Competitive equity and stock options
• 30+ paid days off including bank holidays and company shutdowns
• Private healthcare (virtual and in-person)
• Pension scheme with 8% total contribution
• Free daily breakfast, catered lunch, and snacks in-office

About the Company

Humanoid is on a mission to build the world’s most capable, commercially-scalable, and safe humanoid robots, starting with the HMND-01 Alpha platform.