3+ years of experience in production-level software for real robotic hardware, preferably arms.
Proficient in C++ and/or Python; familiar with Linux and real-time debugging.
Strong understanding of manipulator kinematics, including forward/inverse kinematics and Jacobians.
Experience with a robotics middleware stack (e.g., ROS 2) or a valid reason for not using one.
Hands-on experience integrating robots from vendor SDKs, not just in simulation.
Ability to manage safety-critical code with attention to detail.
Responsibilities
Integrate commercial robotic arms and grippers, including SDKs and calibration.
Evaluate and integrate various operator input devices like VR controllers and haptics.
Develop and refine inverse kinematics and controllers for high DOF robotic arms.
Design systems to handle unavoidable latency and ensure safety limits.
Prototype shared-autonomy features to enhance teleoperation efficiency.
Oversee sensor integration and ensure accurate time synchronization across devices.
Collaborate with a Teleop Platform Engineer on latency budgets and data recording specifications.
Travel to robot sites for hardware setup and debugging as needed.
Benefits
Flexible work environment with opportunities for remote work.
Collaborative team culture focused on innovation and problem-solving.
Access to cutting-edge technology and tools in robotics.
Opportunities for professional development and skill enhancement.
Supportive management that values diverse experiences and perspectives.
Full Job Description
The role
You'll take arms we buy off the shelf and make them safely, precisely controllable by an operator thousands of kilometers away. Many of these arms ship with no teleop stack and no teleop input device, so you'll choose and integrate the operator hardware, write the mapping from human motion to robot motion, and own everything that runs on the robot. You'll also be the person who makes sure every frame of data we record is correctly timestamped and synchronized.
What you'll do
Integrate new commercial arms and grippers: vendor SDKs, drivers, calibration, and a common control interface.
Write and tune inverse kinematics and Cartesian/joint controllers for high DOF robotic arms for teleoperation and manipulation.
Design for latency you can't remove: local safety limits, workspace and collision constraints, graceful handling of dropped or late packets, and safe stop when the link degrades.
Prototype shared-autonomy features (assisted grasping, motion smoothing, task-level primitives) that make long-distance teleop faster and less tiring.
Own sensor integration: cameras, depth, wrist cams, joint state, force/torque. Get time synchronization right across all of them.
Work with a Teleop Platform Engineer on the end-to-end latency budget and on what we record, at what rate, in what format.
Go to robot sites when needed to bring up hardware and debug in person.
What we're looking for
3+ years writing production level software that runs real robot hardware (arms strongly preferred), in industry or a serious research lab.
Strong in C++ and/or Python; comfortable with Linux, real-time considerations, and debugging at the hardware boundary.
Solid grasp of manipulator kinematics: forward/inverse kinematics, Jacobians, singularities, joint limits, and at least one way of doing Cartesian control well.
Experience with a robotics middleware stack (e.g., ROS 2) or a good reason you chose not to use one.
You've integrated a robot from a vendor SDK, not just in simulation.
Comfortable owning safety-relevant code and being careful about it.
Nice to have
Built a teleoperation system before, especially over unreliable or long-distance networks.
Experience with impedance/admittance control, force feedback, or haptics.
Experience with motion planning libraries, simulation, or calibration tooling.
Some networking or systems performance experience: you can read a packet capture and profile a control loop.
Experience with VR/XR
If you don't check every box above but believe you can do the job, we encourage you to apply - we're looking for capability and trajectory, not a perfect checklist match.
