Hardware Systems Architect
Sunnyvale, CA
Hardware Systems Architect Responsibilities
- Own system architectures encompassing compute, displays/optics, EE, sensors, calibration, and firmware from initial conception through prototyping, product development, and production ramp
- Across a portfolio of products, work with ML scientists and SW engineering teams to understand requirements for ML workloads that need to be supported. Drive selection/definition of compute silicon that has appropriate HW accelerators/cores for these workloads
- Lead product architecture for roadmap products, communicate trade-offs, program concerns and technical direction with specific focus on enabling ML-assisted augmented/virtual reality experiences
- Drive the technical scoping and development of new technologies to deliver generation-defining user experiences for future products on the roadmap
- Conduct detailed analysis and trade-off studies for complex systems
- Engage with technology project leads, and various component SMEs in creating the overall system architecture and platform as the starting point for product execution
- Deliver engineering solutions for the most challenging architectural problems in early phase concept development (ensuring translation of marketing requirements to product)
- Communicate technology and product strategy effectively to both internal and external stakeholders
- Guide management and partner teams on long term technology investments
- Create proof-of-concept physical prototypes
- Collaborate with cross-functional teams on blank-slate and early-phase prototyping and product exploration, mapping it to strategic and product questions which need to be resolved
- Create hardware system metrics, KPIs, standards and tools
- Travel up to 15% time domestically and internationally
Minimum Qualifications
- 10+ years of product experience in electrical, systems or computer engineering
- Experience with architecture and systems engineering across electrical, firmware, audio and optical systems
- Hands-on experience with: on-device machine learning frameworks (e.g. TensorFlow, PyTorch etc.), popular NNs and corresponding implementations on relevant hardware, architecture and specifications of compute blocks that are designed to run these models, methodologies for benchmarking compute/memory usage/power consumption and latency on platforms running ML workloads
- Experience with mainstream silicon IPs and SDKs for on-device machine learning, e.g. ARM Mali GPUs, Ethos NPUs, Cadence G/HIFI DSPs, Qualcomm Hexagon MPUs
- Experience integrating and delivering one or more of these technologies: displays, silicon, sensors
- Experience in rapid prototyping and creating proof of concept models to share ideas and improve architectural decisions
- Communication experience working with Product Development engineers, Product Managers and other hardware/software cross-functional teams
- Experience working with partnerships teams and external partners at both a supply base and product level
Preferred Qualifications
- MS/PhD in Electrical Engineering, Computer Engineering other related technical field or equivalent experience
- Experience in firmware development, programming, computer vision, OS, game development, and complex sensing systems
- Experience modeling system power and, using those models, inform trade-offs against key architectural decisions (e.g. battery life, thermals, silicon roadmaps, power load profiles)
- Understanding of memory subsystems including mass storage and filesystem as well as volatile memory and memory management and their impacts on overall system performance
- Experience with AR/VR or systems aimed at capturing/processing/delivering audio visual, IMU or other positional data streams
- Experience creating system models to analyze system compute, latency and communication interface bandwidths across multiple subsystems within a heterogeneous compute environment in order to define the system architecture to meet the overall goals of the product definition
- Experience with influencing silicon vendors roadmaps and SOC features sets with emphasis on using advanced IP in the realms of Machine Learning, Graphics and Imaging
