Software Engineer - Robot Compute Platform

• Own the onboard software platform on NVIDIA Jetson: Real-Time Linux configuration, scheduling, and performance tuning
• Deploy and optimize neural network policies for real-time inference: TensorRT, quantization, zero-copy data paths, strict latency budgets
• Implement and maintain the EtherCAT/CAN master and the joint-level communication with the Motor Controller PCBs
• Integrate sensors: IMU drivers, filtering and time synchronization, cameras and additional sensing as needed
• Build the middleware that moves observations and actions between the bus and the policy at loop rate, deterministically
• Develop logging, replay, and introspection tooling for the whole robot software stack
• Work daily with the RL and Sim2Real engineers on the deployment pipeline, and with embedded on the bus API

• A systems software engineer who thinks in latency budgets and memory copies
• Equally comfortable in CUDA/TensorRT and in a CAN bus trace
• You take full ownership from kernel configuration to inference output, * B.Sc. in Computer Science, Engineering, or a related field
• 8+ years of software engineering with heavy C/C++ focus; deep understanding of modern C++, memory management, and parallelism
• Extensive experience developing and debugging in embedded Linux environments; real-time or low-latency systems experience
• Hands-on experience deploying neural networks on edge platforms (NVIDIA Jetson, TensorRT or equivalent)
• Knowledge of embedded communication protocols: EtherCAT, CAN, SPI, I2C
• Production-grade Python for tooling and pipelines
• Experience with PREEMPT_RT kernels and real-time performance monitoring

Advantages:

• Experience with GPU-accelerated services using zero-copy mechanisms to minimize data transfer latency
• ROS 2 experience
• Background in autonomous driving or edge-AI platforms (e.g., Xpeng, NIO, Horizon Robotics)
• Comfortable communicating technical topics in English with international teams