車電V&V engineer(內湖) - TWPR01276

Key Responsibilities

• Implement V&V process to measure and analyze AI inference performance: accuracy, false positive, false negative, precision, recall, .. etc. Find out and identifying bottlenecks in the AI modules.
• Compliance with test specifications to implement integration test and system test for AI module
• Validate AI inference engine performance and stability across different embedded runtimes (e.g., TensorRT, SNPE, ONNX Runtime).
• Utilize Linux profiling tools and vendor-specific monitors to analyze CPU, GPU, and NPU load during sustained inference tasks.
• Perform stress and reliability testing on inference modules to detect memory leaks, OOM (Out of Memory) issues.
• Develop automated performance benchmarking scripts to ensure consistent validation across different software versions and hardware revisions.
• Collaborate with hardware and software development teams to identify and resolve system-level bottlenecks and hardware-software integration issues.
• Ensure alignment with automotive standards such as: ASPICE (Automotive SPICE), ISO 26262 (Functional Safety), ISO 21448 (SOTIF), OEM-specific requirements (Toyota, Honda, VW, GM, etc.)