Vision Architect, RL Silicon

Meta's Reality Lab Silicon organization is seeking a Vision Architect to define the long-term technical direction for vision and imaging processing capabilities within custom silicon solutions for wearables. In this role, you will establish the architectural vision for computer vision processing units, image signal processors, and related compute subsystems, driving alignment across silicon design, software, and systems teams. You will shape multi-generation silicon roadmaps that enable Meta's wearables at scale, partnering with research, hardware, and product organizations to translate emerging vision workloads into silicon architecture requirements. Responsibilities
Define and own the multi-generation architectural vision for computer vision and image processing subsystems within custom SoCs, including ISP pipelines, computer vision accelerators, and associated memory hierarchies Establish microarchitectural frameworks and design principles for vision processing engines, balancing performance, power, area, and programmability tradeoffs Drive architectural exploration and feasibility studies for next-generation vision workloads, including computer vision inference, video analytics, and real-time image processing on wearable devices Develop and maintain architectural specifications, performance models, and design guidelines that serve as the foundation for RTL implementation and verification teams Collaborate with AI research and software teams to characterize emerging vision workloads and translate computational requirements into silicon architecture decisions Lead cross-functional architecture reviews spanning design, verification, physical implementation, and post-silicon validation to ensure architectural intent is preserved through the full development cycle Define performance benchmarks and silicon KPIs for vision processing subsystems, driving alignment on success criteria across hardware and software stakeholders Evaluate and influence industry trends, emerging standards, and third-party IP strategies relevant to vision and imaging silicon to inform build-versus-buy decisions Mentor other architects and engineers across the silicon organization, establishing architectural best practices and review processes for vision subsystem development Partner with post-silicon and systems teams to validate architectural assumptions, incorporate learnings into future roadmap decisions, and drive first-pass silicon success Minimum Qualifications
Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience 12+ years of experience in silicon architecture, algorithm development, or hardware design with a focus on computer vision, imaging, or video processing subsystems Experience developing cycle-accurate or transaction-level performance models for vision or imaging processing engines Experience in driving architecture definition across full silicon development cycles, from concept through post-silicon validation, for wearable applications Experience collaborating with AI, computer vision, or machine learning teams to map algorithmic workloads onto custom silicon architectures Experience architecting computer vision and imaging processing subsystems for AI inference acceleration, including integration with neural network accelerators or tensor processing pipelines Experience influencing multi-generation silicon roadmaps through written architectural proposals and executive-level technical communication Experience with software-hardware co-design and developing models that interact with low-level system components Experience with designing image processing or computer vision algorithms optimized for specialized hardware such as GPUs, DSPs, or custom ASICs