Lead Edge AI SoC Runtime Architect

Velaura; Santa Clara (CA); On-site; USD 200,000 - 500,000; Full time

Job summary

Velaura seeks a Principal AI SoC Runtime Software Architect to own the software architecture for turning Velaura’s heterogeneous AI SoC into a coherent, high‑performance execution platform. You will define end‑to‑end runtime across sensor ingest, preprocessing, AI inference, postprocessing, and delivery to robotics and edge AI apps.

The role requires hands‑on leadership across runtime, kernel, driver, firmware, multimedia, and SDK development, with a focus on low latency, data movement, and

Qualifications

• Extensive experience designing production runtime systems or similar performance-critical software.
• Strong C/C++ programming and hands-on experience across user-space APIs, kernels, firmware, and hardware interfaces.
• Deep understanding of heterogeneous and asynchronous execution, queues, dependencies, synchronization, and resource management.
• Knowledge of device memory, DMA, IOMMU/SMMU, cache coherency, and shared buffers across kernel/user space.
• Experience optimizing end-to-end data movement across multiple hardware engines.
• Designing stable runtime APIs with versioning, error handling, diagnostics, and recovery.
• Ability to debug cross-layer correctness and performance issues using data-driven methods, profiling, and tracing.
• Technical leadership across components and teams, translating requirements into architectures and validation plans.

Responsibilities

• Define the SoC-wide execution model coordinating workloads across heterogeneous compute and media engines.
• Set architecture and direction for the AI inference runtime and interfaces to applications and the SDK.
• Own end-to-end dataflow for sensor-to-application pipelines across camera, preprocessing, inference, and postprocessing.
• Define memory and buffer-sharing architecture across user space, kernel, and hardware engines with clear ownership and lifecycles.
• Define division of responsibility and interface contracts among runtime, kernel drivers, firmware, and hardware engines.
• Partner with compiler team to define compiler–runtime contract and ensure metadata for loading, validating, and profiling.
• Establish system-wide observability and performance architecture for latency, throughput, bandwidth, and power goals.
• Define resilience and validation architecture including fault containment, recovery, and acceptance criteria.

Skills

C/C++ programming

Runtime systems

Linux kernel

DMA and shared memory

Performance optimization

Debugging and profiling

Technical leadership

Tools

ONNX Runtime

TensorRT

GStreamer

Job description

Velaura seeks a Principal AI SoC Runtime Software Architect to own the software architecture for turning Velaura’s heterogeneous AI SoC into a coherent, high‑performance execution platform. You will define end‑to‑end runtime across sensor ingest, preprocessing, AI inference, postprocessing, and delivery to robotics and edge AI apps.

The role requires hands‑on leadership across runtime, kernel, driver, firmware, multimedia, and SDK development, with a focus on low latency, data movement, and