Robotics Software Engineer - Perception

About the role

We are looking for a Robotics Software Engineer specialized in perception to own the autonomy stack of our agricultural robots, from perception and state estimation through navigation and motion control, and to validate the software on real hardware in the field.

Key responsibilities

• Own the localization and mapping stack for GPS‑degraded, unstructured environments.
• Develop the perception layer that fuses multimodal sensor data into reliable models of obstacles, free space and configuration space.
• Design navigation and motion behaviours, including planning, control, smoothing and recovery strategies.
• Implement real‑time state estimation and sensor fusion under noise and uncertainty.
• Maintain a clean, modular and testable codebase to enable rapid iteration between simulation and hardware.
• Develop and sustain simulation environments for end‑to‑end autonomy testing before deployment.
• Contribute to engineering standards such as reproducible development environments, linting, automated testing and profiling.
• Participate in recurring lab and field sessions, define objectives, run trials, log metrics and feed findings back into the backlog.
• Collaborate with a multidisciplinary team committed to building general‑purpose physical AI.

Required profile

• 2+ years of robotics software development experience and a Master’s degree in Robotics, Computer Science, Electrical/Mechanical Engineering or a related field.
• Proven ability to take autonomy projects from perception and state estimation through navigation and motion control to real‑hardware validation.
• Strong understanding of localization, mapping and sensor‑fusion techniques in noisy, real‑world conditions.
• Experience with perception from range and vision sensors, including filtering, segmentation, registration and multimodal data fusion.
• Familiarity with modern robotics middleware and real‑time inter‑process communication.
• Comfort with simulation‑driven development and containerised, version‑controlled workflows.
• Commitment to data and code quality, rigorous evaluation and meticulous attention to detail.
• Self‑starter attitude with the ability to debug on real robots in imperfect field conditions.

Required skills

• C++
• Python
• Robotics middleware (e.g., ROS)
• Sensor fusion and state estimation
• Localization and mapping
• Perception processing for range and vision sensors
• Simulation environments
• Containerisation (Docker)
• Version control (Git)
• Real‑time inter‑process communication