Lead Software Engineer - Dexterous Manipulation

Apptronik Apptronik · Robotics · HQ · Advanced Technologies

Lead Software Engineer for dexterous manipulation in robotics, focusing on developing learning-based control algorithms for human-like precision in robotic hands. This role bridges research and production, involving architectural definition, sim-to-real strategy, and influencing hardware design for scalable deployment of AI-powered robots.

What you'd actually do

  1. Serve as the technical authority for dexterous manipulation. Create the long-term technical roadmap, ensuring hand control and multi-fingered coordination capabilities outpace industry standards.
  2. Design and enforce the foundational software frameworks for manipulation. Own the decision-making process for balancing autonomous logic with high-fidelity teleoperation, ensuring the architecture is scalable for future hardware generations.
  3. Perform and direct the integration of state-of-the-art research. Select and deploy the specific learning-based policies and vision-integrated systems that will define system physical capabilities.
  4. Lead the strategy for high-fidelity simulation. Set the standards for success in virtual environments to optimize policy transitions to physical fleet hardware.
  5. Oversee the transition from experimental research to fleet-wide deployment. Ensure performance and reliability of C++/Python code running on production-level assets.

Skills

Required

  • Dexterous Manipulation
  • multi-fingered hand control
  • grasp planning
  • in-hand manipulation
  • Advanced Control & Learning
  • learning-based control for robotics
  • flow/diffusion-based visiomotor policies
  • reinforcement learning
  • reward modeling
  • Python
  • real-time robotic software stacks
  • Simulation Environments
  • IsaacSim
  • MuJoCo
  • Drake

Nice to have

  • Robotic Kinematics
  • spatial transformations
  • Jacobian-based control
  • constrained optimization
  • Teleoperation
  • VR/haptic interfaces
  • retargeting algorithms
  • Tactile Sensing
  • haptic feedback
  • Computer Vision
  • 6D pose estimation
  • point cloud processing
  • visual-servoing
  • Hardware Bring-up
  • end-effector calibration

What the JD emphasized

  • strong track record of successful hardware deployment
  • proven track record of taking complex algorithms from a research/simulation environment and successfully deploying them on physical hardware

Other signals

  • embodied AI
  • dexterous manipulation
  • learning-based control
  • sim-to-real
  • production software