Staff AI Application Engineer (Automotive SoC)
Renesas Electronics4 days ago
Seoul, Korea, SouthStaff+
Responsibilities
- Enable and deploy BEV, object detection, segmentation, and classification models on Gen4/5 automotive SoC platforms with CNNIP, DSP, NPU, and HWA acceleration.
- Analyze latency, throughput, multi-core scaling, memory bandwidth, scheduling, and operator-mapping bottlenecks.
- Support PTQ, QAT collaboration, operator fusion, graph optimization, execution partitioning, and quantization-accuracy mitigation.
- Integrate models into Linux and QNX embedded runtimes and debug offloading, memory allocation, IPMMU, data-transfer, and synchronization issues.
- Validate AI workloads on target boards and SIL/HIL simulators.
- Support ONNX-based compiler and runtime workflows, including graph inspection, graph modification, model segmentation, execution control, and QDQ models.
- Develop or maintain tools and scripts for model validation, benchmarking, and customer workflows.
- Support customer evaluations, proofs of concept, demos, technical guidance, documentation, issue tracking, reports, and release validation.
Requirements
- Bachelor’s or master’s degree in computer science, electrical engineering, or embedded systems, or experience in embedded systems.
- Solid understanding of deep learning fundamentals and inference pipelines.
- Hands-on experience with PyTorch, ONNX, or ONNX Runtime.
- Strong Python programming skills; C/C++ knowledge is a plus.
- Familiarity with embedded systems and debugging tools.
- Ability to analyze latency, throughput, and hardware-utilization performance metrics.
- Preferred experience includes 1–3 years in embedded systems or AI-related development.
- Preferred qualifications include model training, fine-tuning or evaluation; computer vision, automotive or robotics use cases; embedded NPU/DSP/GPU/CPU inference optimization; quantization and INT8 calibration; automotive SoCs or QNX; memory hierarchy, DMA, and multi-core scheduling.
- Customer or application-engineering support, automotive AI or ADAS perception knowledge, internal tooling or documentation, and ONNX graph or intermediate-representation debugging are nice to have.
Benefits
- Regular permanent employment; remote work is not available.
- Opportunity to work on automotive AI platforms across the AI lifecycle from models to silicon.
- Collaboration with compiler, runtime, and hardware architecture teams and exposure to customer deployments.
- Competitive benefits package; further details are provided during the hiring process.
About Renesas Electronics
Renesas Electronics designs and sells microcontrollers, analog and power ICs, and connectivity solutions for automotive, industrial, infrastructure and IoT manufacturers. Its portfolio spans embedded processing, power management and ADAS-ready components, sold directly and through channels to OEMs and Tier-1 suppliers, with software and development tools to integrate them. Founded in 2010 and headquartered in Tokyo, it is a public company on the Tokyo Stock Exchange and has expanded through acquisitions including Intersil and Dialog Semiconductor.