Robotics & Physical AI Engineer

The work

Turn a customer’s task into a working robotics prototype with the team. Follow the full path from sensing and planning to motion, then measure what happened on the hardware. Help define the operating limits, investigate failures, and make the next iteration better.

Professional experience is not required. Strong web application developers are welcome, even if robotics and hardware are new to you. Show us what you can build, explain how your code works, and bring a willingness to learn the physical side of the product.

What we look for

• Strong implementation skills: build and debug a working application or device, explain your design choices, and improve it when something fails. Personal projects, study projects, and open-source work count.
• Web application skills: if your background is in web development, show us a frontend connected to an API and a backend or data store. Be ready to discuss state, data flow, error handling, tests, and the code you wrote. React/TypeScript, Node.js, Python, or comparable technologies are welcome.
• A practical learning habit: use Git, read documentation, investigate problems with logs and repeatable tests, and learn unfamiliar tools. Prior experience with ROS 2, C/C++, or electronics is helpful but not required for applicants with strong web application skills.
• Care with physical systems: learn the equipment’s limits and agree test procedures with the team before powering or moving hardware. Ask for help when a task is outside your knowledge.

Where you can contribute and grow

Start from strong web application skills or an existing hardware or robotics specialty. The examples below come from our recent engineering articles and show what you can learn with the team; you do not need to know every technology before applying.

• Web applications and robot communication: build operator interfaces, device APIs, and telemetry dashboards. Follow commands and measured state across USB/UART, a smart-servo bus, or CAN/CAN-FD; distinguish these links from SPI and I²C connections to nearby sensors.
• Mechanics and actuation: read CAD and assembly drawings; relate degrees of freedom (DOF), payload, reach, and torque–speed limits to the task. Compare stepper motors and brushless motors (BLDC/PMSM), gear reduction, and belt or cable drives. Check backlash, encoder placement, and joint calibration.
• Electronics and power: read datasheets and schematics, check voltage and pinouts, and distinguish GPIO signals from motor power. Learn how PWM signals, H-bridge drivers, and field-oriented control (FOC) drive different motor types. Understand DC/DC conversion, current limits, and regenerative energy when a motor slows down.
• Embedded firmware and real-time control: use C/C++ on microcontrollers such as Pico 2, ESP32, or STM32. With FreeRTOS or Zephyr, learn task scheduling, interrupt service routines (ISRs), queues, semaphores, mutexes, and priority inversion. Measure control-loop latency and jitter; understand where Linux PREEMPT_RT and local microcontroller control fit.
• Robotics software, simulation, and learning: connect ROS 2/DDS communication, coordinate frames, and inverse kinematics to robot motion. Our FORTE arm experiments use MuJoCo for physics, Gymnasium for the learning environment, and PyTorch for behavioral cloning from demonstrations. Learn to evaluate recorded trajectories and investigate the gap between simulation and hardware (sim-to-real).
• Device and command security: understand device/topic permissions in MQTT and DDS-Security for ROS 2. Learn secure boot, signed OTA firmware updates, and recovery. Check command expiry and replay protection, and validate AI-generated plans before they can trigger motion.

Useful additional experience

• Oscilloscopes or logic analyzers, PCB design, soldering, or building serviceable cable harnesses.
• Tuning PID feedback control with encoder data, or measuring the full sensor-to-actuator delay under load.
• Camera calibration or investigating sim-to-real failures with reproducible logs, datasets, and test conditions.
• Diagnosing bus faults, power dips, or motor overheating, and checking that recovery works on the hardware.

What to show us

Share a web app, GitHub repository, demo, robot project, or short technical write-up. Explain what you built yourself, one problem you solved, and how you checked the result. A working web application is welcome evidence; a robot portfolio or previous job is not required. Include photos, video, or CAD if they help explain your work.

Read the engineering behind the role

Our recent articles show the topics we are exploring. They include our prototype work and wider technical research; a topic appearing in an article does not mean it is already deployed in a customer system.

• Physical AI hardware: motors, actuators, and validation
• Robot electronics: circuits, controllers, and power
• RTOS architecture: scheduling, interrupts, and robot control
• Physical AI security: devices, commands, and recovery
• Our FORTE arm: MuJoCo, behavioral cloning, and hardware development