Principal AI SoC Runtime Software Architect

What you'll do

• Define the SoC-wide execution, dataflow, and memory-sharing architectures across heterogeneous compute and media engines.
• Set the AI inference runtime architecture and its interfaces with applications, SDK components, drivers, firmware, and compilers.
• Lead observability, performance optimization, resilience, and validation across software and hardware layers.

What they ask for

• Extensive experience designing and building production runtime systems, embedded middleware, multimedia frameworks, or other performance-critical systems software.
• Strong C/C++ skills and hands-on architecture experience spanning application APIs, user-space libraries, and low-level interfaces.
• Strong understanding of heterogeneous execution, device memory, DMA, IOMMU/SMMU, cache coherency, shared buffers, and kernel/user-space memory interfaces.
• Experience optimizing end-to-end data movement and execution across hardware engines; designing stable runtime APIs; debugging cross-layer issues; and providing technical leadership.

Nice to have

• Experience developing AI inference runtimes or execution providers, such as ONNX Runtime, OpenVINO, or TensorFlow Lite delegates.
• Linux kernel development, or experience with robotics, edge AI, camera pipelines, embedded Linux SDKs, or compiler/runtime integration.

Pay

USD 200,000–300,000 base salary

Tools

C/C++, ONNX Runtime, TensorRT, OpenVINO, TensorFlow Lite, Qualcomm QNN/SNPE, TVM, Linux kernel

AI focus

Architects and leads an AI inference runtime for a heterogeneous AI SoC, integrating compiled models and coordinating inference with sensor, preprocessing, and postprocessing pipelines.