Integration Engineer - Perception and Platform

AMD AMD · Semiconductors · San Jose, CA · Engineering

Senior Software Engineer on the Physical AI team, focusing on deploying and optimizing AI inference workloads and perception pipelines on AMD adaptive computing platforms for robotics, automotive, and intelligent edge applications. This role involves working with AI models, hardware acceleration, and embedded systems to build intelligent machines.

What you'd actually do

  1. Design, develop, and optimize software for Physical AI applications targeting AMD adaptive computing platforms.
  2. Deploy and optimize AI inference workloads using Vitis AI and related toolchains.
  3. Develop advanced perception pipelines for robotics, automotive, and intelligent edge applications.
  4. Integrate AI models into high-performance embedded and edge computing environments.
  5. Analyze and optimize end-to-end system performance, latency, throughput, and power consumption.

Skills

Required

  • C++
  • Python
  • Linux
  • deploying and optimizing machine learning or deep learning models for edge devices
  • AI inference pipelines and performance optimization techniques
  • computer vision, perception systems, or sensor processing applications
  • embedded systems and heterogeneous computing architectures
  • profiling and optimizing software performance across CPU, GPU, NPU, FPGA, or accelerator-based systems
  • debugging, problem-solving, and system integration skills
  • written and verbal communication skills
  • multidisciplinary engineering teams

Nice to have

  • AMD Vitis AI or equivalent AI deployment frameworks
  • AMD Versal AI Edge, Zynq UltraScale+, Kria, or other adaptive computing platforms
  • FPGA acceleration and hardware/software co-design methodologies
  • ONNX, PyTorch, TensorFlow, or similar frameworks
  • quantization, model optimization, pruning, and edge deployment techniques
  • ROS 2 and robotics software architectures
  • perception systems including camera, radar, lidar, and sensor fusion pipelines
  • graphics, visualization, or accelerated rendering technologies
  • OpenGL, Vulkan, Wayland, or embedded graphics frameworks
  • Linux kernel, device drivers, or low-level platform software development
  • distributed AI, edge-to-cloud architectures, or heterogeneous computing systems
  • robotics, automotive, aerospace, industrial automation, or other intelligent edge domains

What the JD emphasized

  • deploy and optimize machine learning or deep learning models for edge devices
  • AI inference pipelines and performance optimization techniques
  • computer vision, perception systems, or sensor processing applications
  • embedded systems and heterogeneous computing architectures
  • profiling and optimizing software performance across CPU, GPU, NPU, FPGA, or accelerator-based systems

Other signals

  • Deploy and optimize AI inference workloads
  • Develop advanced perception pipelines
  • Integrate AI models into high-performance embedded and edge computing environments
  • Analyze and optimize end-to-end system performance, latency, throughput, and power consumption
  • Evaluate emerging AI models, frameworks, and deployment methodologies