Machine Learning Staff Software Engineer, Pixel Camera

Overview In this role you will design and implement a real-time ML foundation that turns raw camera and sensor data into actionable signals on-device. You will optimize models for NPU, GPU, and DSP, collaborating with Tensor silicon and compiler teams to craft hardware-aware architectures and custom kernels. You own end-to-end model life cycles—from research prototypes to production on hundreds of millions of devices—leveraging JAX and TensorFlow pipelines and integrating C++ into the camera stack. You’ll partner across product, UX, software, and hardware teams to translate roadmaps into feasible device-level designs, set best practices, and shape the long-term ML roadmap. Compensation / Benefitsbonus target (20%)equitybenefitscompetitive compensation (US: 207000-300000) ResponsibilitiesDesign and implement real-time on-device ML foundation for signal extraction from camera/sensor inputOptimize models for on-device accelerators (NPU, GPU, DSP) including quantization and hardware-aware architecture designDevelop and maintain end-to-end model lifecycle from prototype to production on mobile/embedded devicesBuild training pipelines using JAX and TensorFlow; integrate C++ into the camera pipelineCollaborate with product, UX, software, and hardware teams to translate roadmaps into device-constrained designsEstablish best practices for ML development, deployment, and evaluation; define model quality metrics and contribute to the technology roadmap Key requirementsBachelor's degree in any field or equivalent practical experience8 years of experience in software design and architectureExperience with C, C++, machine learning, and embedded systemsExperience with machine learning algorithms, architectures, and researchCollaboration across cross-functional teamsLeadership and initiativeAdaptability and problem-solving in fast-paced environmentsC, C++Machine learning algorithms and architecturesEmbedded systems experience