Robotics SW Engineer

Role Overview

The WoRV team builds Mobile Manipulation models that solve driving + attachment control together in brownfield environments.

This role bridges the AI autonomy stack and real robot hardware, ensuring that software decisions translate into safe and reliable actions in the field.

We work across Simulation, Hardware, and the AI autonomy stack (VLA, Motion Planning, Localization, Control). Some capabilities are implemented as explicit modules; others are extended as learned policies or VLA models. The Robotics SW Engineer integrates these diverse technology elements into systems that run on real equipment, implementing sensors, actuators, control loops, state management, and fallback logic at product level.

The core goal is not to make a single PoC succeed, but to build Safe, Reliable, and Scalable robot software capabilities that can be applied repeatedly across multiple customers and field sites.

Responsibilities

• Design and develop robot software systems including driving, attachment control, sensors, actuators, and safety logic.
• Implement real-time control loops, state management, command interfaces, and mission execution flows for Mobile Manipulation products.
• Integrate Localization, Motion Planning, Control, Perception, and VLA/learned policies to run reliably on real robots.
• Implement and optimize sensor-actuator communication, data synchronization, control command delivery, and status monitoring architectures.
• Design fallback logic and safety validation for brownfield noisy sensors, partial observability, exceptions, and edge cases.
• Analyze field logs and failure cases to improve control stability, latency, robustness, and maintainability.
• Collaborate with PM / AI / Simulation / Data teams to convert customer pain points into reusable product modules.
• Accumulate reusable robot control modules, validation procedures, and operational tools through the PoC → Pilot → Production process.

Qualifications

• Experience implementing and debugging real systems in C/C++
• Ability to develop, build, deploy, analyze logs, and debug performance in Linux environments
• Experience controlling or integrating at least one of: robots, vehicles, equipment, sensors, or actuators
• Basic understanding of at least one of: PID, kinematics, dynamics, state estimation, trajectory tracking, or motion control
• Interest in system design involving sensor data, control commands, state machines, fault handling, and safety checks
• Ability to structurally analyze failure causes from logs, video, sensor data, and operational results
• Ability to quickly learn new hardware and software stacks, and pivot to simpler, more reliable approaches when needed

Preferred Qualifications

• ROS2-based robot system development, node composition, topic/service/action design, and launch/config management
• RTOS, MCU, ARM, embedded Linux, or other embedded/real-time system development experience
• Experience implementing sensor-actuator communication protocols: CAN, UART, I2C, SPI, EtherCAT, UDP/TCP
• Experience applying PID, MPC, WBC, trajectory optimization, model-based control, or learning-based control
• Experience controlling real equipment: mobile robots, manipulators, attachments, AGV/AMR, vehicles, agricultural or construction machinery
• Experience with brownfield environments, noisy sensors, partial observability, and edge cases
• Experience with edge devices, on-device inference, real-time inference, or low-latency system optimization

Who We're Looking For

• Someone who knows the difference between an algorithm working well and real equipment moving safely
• Someone who asks "which technology fits this problem?" first, and can design simple rule-based and AI-based methods together when needed
• Someone who analyzes robot failures not as simple bugs, but as system problems involving sensors, control, environment, and operational conditions
• Someone who wants to turn lab-proven capabilities into repeatable product capabilities for customer field sites
• Someone who communicates quickly across hardware, AI, data, and PM teams to solve problems end-to-end
• Someone who wants to leave behind reusable modules and validation frameworks across multiple field sites, not just local optima