Principal AI SoC Runtime Software Architect

Brahma Consulting Group; Santa Clara (CA); Sur place; USD 260 000 - 330 000; Plein temps

Résumé du poste

Brahma Consulting Group partners with a high-growth startup building ultra-low-power AI SoCs for robotics and real-world physical AI. We seek a Principal AI SoC Runtime Software Architect to own the runtime architecture at the silicon platform’s core, shaping data flow from sensors to robotics applications.

This architect-level role offers hands-on leadership to design, optimize, and integrate runtime, kernel drivers, and multimedia software, with a clear path to build and lead a dedicated

Qualifications

• 8–10+ years in low-level systems software or architecture
• On-chip/on-device runtime experience on embedded SoCs
• Deep knowledge of DMA, cache coherency and heterogeneous compute coordination
• Experience with camera/imaging pipelines (V4L2, ISP, OpenCL, Android camera stack)

Responsabilités

• Architect the embedded runtime that manages data flow from sensors through image processing, NPU inference, to robotics applications.
• Orchestrate heterogeneous compute across CPUs, DSPs, NPUs, ISP with low-latency synchronization.
• Define robust interfaces between runtime, Linux kernel drivers, firmware, compilers, and external SDKs.
• Cross-layer performance tuning to reduce latency, improve throughput, memory bandwidth, and power.
• Provide technical leadership to runtime, kernel driver, and multimedia teams and scale the software organization.

Connaissances

Senior results leadership

Embedded runtime experience

Hardware & arch

Multimedia pipelines

C/C++/Embedded Linux

Technical leadership

Formation

Bachelor's degree in Electrical/Computer Engineering or CS

Outils

Linux kernel drivers

V4L2 / GStreamer

Description du poste

We are partnering with a high-growth, heavily funded ($300M+ raised) Physical AI and Edge Compute startup founded by semiconductor industry veterans. Backed by top-tier venture firms, they are building revolutionary, ultra-low-power AI SoCs designed to power the next generation of robotics, smart machines, and real-world physical AI applications.

The Opportunity

We are seeking a Principal AI SoC Runtime Software Architect to serve as a key early technical leader. In this hands‑on, high‑impact role, you will own the runtime architecture at the center of the silicon platform—defining and building the software that orchestrates end‑to‑end data flow from sensor ingest all the way to robotics application layers.

This is an architect‑level position with a clear path to build and lead a dedicated team of runtime, driver, and systems software engineers around your architecture.

What You Will Do

• Architect the Runtime Pipeline: Design and build the embedded runtime that manages data flow from sensors (camera, audio, IMU) through image processing, NPU inference, and postprocessing up to robotics applications.
• Orchestrate Heterogeneous Compute: Optimize execution across heterogeneous engines (CPUs, ISPs, DSPs, NPUs) using DMA, cache coherency, shared zero‑copy buffers, and low‑latency synchronization.
• Define System Interfaces: Establish robust APIs and interfaces connecting the runtime with Linux kernel drivers, low‑level firmware, compilers, and external SDKs.
• Cross‑Layer Performance Tuning: Debug and optimize end‑to‑end latency, throughput, memory bandwidth, and power consumption across hardware and software boundaries.
• Technical Leadership: Lead and mentor engineering initiatives across runtime, kernel driver, and multimedia teams while scaling the software organization.

What We Are Looking For

• Seniority: 8–10+ years of experience in low‑level systems software, embedded runtime development, or systems architecture.
• Embedded / On‑Chip Runtime Experience: Proven track record of shipping production runtime software on an embedded SoC (Note: We are specifically looking for on‑device/on‑chip experience, not data center or rack‑to‑rack systems).
• Hardware & Computer Architecture: Deep domain knowledge in DMA, cache coherency, IOMMU/SMMU, and heterogeneous compute coordination (CPU, NPU, DSP, ISP).
• Multimedia & Vision Pipelines: Experience working with camera/imaging pipelines (e.g., V4L2, GStreamer, ISP, OpenCL, Android camera stack).
• Core Tech Stack: Proficiency in C, C++, Embedded Linux, and Linux kernel drivers.
• Technical Leadership: Prior experience leading runtime, driver, or multimedia software projects with multiple engineers.

Nice‑to‑Haves

• Experience with ML inference runtimes or compilers (e.g., ONNX Runtime, TensorRT, OpenVINO, TensorFlow Lite, Qualcomm QNN/SNPE, TVM).
• Exposure to robotics middleware (e.g., ROS 2).
• Degree in Electrical Engineering, Computer Engineering, or Computer Science (or equivalent practical experience).