Senior Software Engineer, Navigation

About The Role

Join the Autonomy team as a Senior Software Engineer focused on Navigation. You will be a core contributor, driving the design and deployment of real-time motion planning and navigation systems that empower our humanoid robots to operate robustly and autonomously in complex logistics and manufacturing environments. This is a high-impact role essential to scaling our commercial deployments, improving the efficiency of autonomous loco-manipulation behaviors, and achieving software readiness for the next generation robot platforms.

About The Work

• Design, implement, and deploy 3D motion planning algorithms for locomotion, with an emphasis on whole-body collision-aware motion execution in real-time.
• Own the core components of our navigation stack, specifically the local planning maps, terrain models, and grid map representations used for path and motion planning with collision avoidance.
• Advance our locomotion capabilities by developing and maintaining a 3D footstep path planner aiming to significantly reduce navigation cycle times.
• Define and implement the necessary navigation features and route planning algorithms to enable coordinated movement and resource sharing within multi-agent robot fleets.
• Drive the maturity of our release processes by designing, implementing, and maintaining robust regression testing pipelines for motion planning and navigation modules.
• Collaborate with the AI and Controls teams to integrate locomotion behaviors with RL policies and support whole-body control.
• Integrate and debug planning algorithms on real-world hardware, owning the transition from simulation environments (e.g., Gazebo, MuJoCo, Isaac Sim) to physical robots.

About You

• 5+ years of professional experience in robotics, specifically developing and deploying real-time navigation and motion planning systems for autonomous mobile platforms (humanoids, quadrupeds, autonomous vehicles).
• Expertise in 3D/volumetric map representations (e.g., octomaps, voxel grids) for local path planning and collision avoidance.
• Deep technical understanding of locomotion-specific path and motion planning algorithms, including sampling-based planners (RRT/PRM), optimization-based methods (MPC/LQR), and hybrid A*.
• Expert proficiency in modern C++ (C++17/20), with a proven track record of writing high-performance, multithreaded code for robotics applications.
• Experience with common robotics frameworks (e.g., ROS/ROS2, DDS) and hands-on experience with modern optimization libraries relevant to motion planning (e.g., Ceres, IPOPT, OSQP).
• Proven ability to systematically test and debug systems on physical robots, and integrate perceived environment data (LiDAR, camera, depth sensing) into the planner.

Bonus Qualifications

• Experience training and deploying Reinforcement Learning (RL) agents for complex locomotion behaviors.
• Hands-on experience implementing Model Predictive Control (MPC) or similar optimization-based control techniques for dynamic robot locomotion.
• Familiarity with perception pipelines and the integration of perceived environment data into the planning stack.
• Experience with GPU-accelerated spatial data structures (e.g., NVBlox, specialized CUDA implementations) for high-throughput, low-latency map updates and querying.
• Strong foundational knowledge of robot kinematics, dynamics, controls, and state estimation (e.g., EKF, particle filters).
• Experience with multi-robot coordination/route planning and abiding by boundary constraints in a workcell map.
• Experience with multi-robot mapping and localization, including map persistence and sharing capabilities.
• Publications in top-tier robotics conferences (ICRA, RSS, IROS, CoRL).

This a hybrid position based out of one of our Salem, Pittsburgh, or Fremont offices.

The final salary offered to a successful candidate will be dependent on several factors that may include but are not limited to: market location, job-related knowledge, skills, and experience. This range may change based on geographical location and may be modified in the future.