Robotics Platform Jetson Integration Engineer

Rivr seeks a mid-level Robotics Platform Jetson Integration Engineer in Zürich to optimize AI workloads on embedded hardware for autonomous delivery robots.

Amazon Rivr develops wheeled-legged robots for last-mile delivery in complex urban settings. This role focuses on integrating and optimizing AI workloads on NVIDIA Jetson platforms to ensure reliable, low-latency performance for autonomous navigation and delivery tasks.

Responsibilities

• Design and execute computing strategies that distribute tasks across the Jetson's CPU, GPU, and DLA to maximize unified memory efficiency and reduce data transfer delays.
• Oversee the deployment and tuning of reinforcement learning models and neural networks, prioritizing fast execution and consistent behavior in changing environments.
• Utilize profiling utilities to monitor system performance and manage Jetson resources, ensuring stable operation of the software stack while identifying and resolving bottlenecks.
• Advise on architectural decisions that enable the platform to scale effectively without compromising performance.

Requirements

• Bachelor's or Master's degree in Computer Science, Electrical Engineering, or Robotics.
• Minimum of one year of practical experience handling resource allocation and prioritization between CPU and GPU workloads.
• Proficiency in navigating large C++ and Python codebases within robotics applications.
• Familiarity with machine learning frameworks including CUDA and TensorRT.
• Proven track record of deploying complex software and reinforcement learning policies on embedded Linux systems, particularly Jetson devices.
• Strong diagnostic skills to troubleshoot and resolve issues when software or RL policies fail on embedded hardware.

Nice to have

• Demonstrated capability in debugging intricate software on embedded hardware, specifically Jetson platforms.
• Comprehensive knowledge of Linux kernel internals, particularly scheduling and memory management, relevant to high-performance embedded systems.
• Experience with NVIDIA DeepStream or Isaac ROS for hardware-accelerated perception pipelines.
• Practical expertise in model optimization techniques such as precision calibration and layer fusion to enhance hardware-accelerated throughput.
• Experience implementing Continuous Integration/Continuous Deployment (CI/CD) practices for embedded systems.

About the company

Amazon Rivr, an ETH Zurich spin-off acquired by Amazon in March 2026, pioneers Physical AI through real-world doorstep delivery. The company builds wheeled-legged robots capable of navigating complex, unstructured environments like stairs, gates, and uneven urban terrain. Rivr is committed to fostering a diverse and inclusive team where collaboration is valued, requiring in-person presence in their office locations to solve real customer problems using rich operational data.

• Focus on solving real customer problems in the real world to achieve general physical intelligence.
• Commitment to building a diverse and inclusive team that values every perspective.
• Collaborative work environment requiring in-person presence in office locations.