All Three Roles Require

• 7+ years of experience in software development for Robotics
• Strong ROS 2 experience — writing nodes, managing topics/services/actions, structuring a ROS-based system
• Systems integration mindset: interfacing with sensors, working with hardware drivers, connecting real-time controllers to higher-level software
• Proficiency in Python; C++ comfort (you don't need to be a C++ expert, but you need to be able to read, extend, and improve existing C++ code)
• Experience running software on NVIDIA embedded compute (Jetson Orin/Xavier or similar)
• Able to work independently in a fast-moving, greenfield environment

Role 1 — Edge AI & Perception Engineering

You're the engineer who takes a model that a scientist trained and makes it run efficiently, and reliably, on embedded hardware. You're not a researcher, but you understand the stack from PyTorch model to TensorRT-optimized inference on the edge. You’ll own the perception pipeline: getting sensor data in, running models, and getting usable output to the rest of the system

Role 2 — SLAM, CV & Spatial Perception

You’re a perception and localization engineer at home with 3D sensing, point clouds, and understanding where a robot is in the world. You’ll own the spatial perception side of the stack, SLAM, mapping, and the geometric understanding the robot needs to navigate and interact with its environment.

Role 3 — Platform Architecture & Software Quality

The robotics software framework exists and works, but it was built by scientists moving fast — it’s not as maintainable, extensible, or performant as it needs to be. You’re the engineer who looks at that codebase and knows how to make it better: cleaner architecture, better abstractions, performance improvements, Python-to-C++ migration where it counts. You care about code quality and know how to improve a system without breaking it.