Principal Engineer, SoC NPU Architect
About this role
Engineers who thrive on deep hardware-software co-design and complex neural processing unit architecture will find a natural home in the Samsung SoC Architecture Lab in Mountain View. The group shapes foundational technology for upcoming Galaxy smartphones, internet of things hardware, and flagship devices through targeted research and early-stage exploration of key IP blocks. Daily work centers on proposing micro-architectural features, building cycle-accurate performance models, and mapping advanced machine learning use cases to silicon to improve overall performance, power, and area.
Leadership values individuals who can balance high-level architectural vision with hands-on simulation and cross-functional collaboration across driver, operating system, and hardware teams. The group exists to drive the corporation's artificial intelligence strategy forward by defining high-performance system-on-chip architectures that impact hundreds of millions of users globally. Candidates who combine extensive processor development background with strong analytical skills and a collaborative mindset will succeed in guiding these next-generation hardware designs from concept to reality.
What you will do
- Propose innovative neural processing unit architectural and micro-architectural features to boost performance, power, and area on targeted workloads for upcoming system-on-chip designs.
- Perform high-level performance modeling, simulation, and analysis covering applications, benchmarks, and complex use cases.
- Develop cycle-accurate and analytical performance models and correlate them directly with silicon results.
- Deliver architectural and microarchitectural proposals and specifications to the design team while communicating effectively with technical leadership and peers.
- Analyze benchmarks, application workloads, and simulation results to drive architecture and microarchitecture optimizations.
- Map artificial intelligence use cases to the neural processing unit while optimizing and co-designing hardware and software for low power and high performance.
- Collaborate across multiple teams to bring architecture proposals to fruition across the system-on-chip, driver, operating system, and overall system.
What you need
- Hold a Bachelor's, Master's, or PhD degree in Electrical Engineering, Computer Science, a related science field, or possess an equivalent combination of education, training, and experience.
- Bring 16 or more years of professional experience specifically focused on neural processing unit architecture and development.
- Demonstrate hands-on experience in developing cycle-accurate performance and functional models.
- Possess good knowledge and practical experience working with convolutional neural networks and large language models.
- Exhibit proficient programming skills utilizing Python, SystemC, and/or C++.
- Show documented experience working with machine learning frameworks such as TensorFlow, PyTorch, and ONNX.
- Bring familiarity with the microarchitecture of system-on-chip designs, interface subsystems, and logic modules.
Nice to have
- Experience in performance and power modeling using Platform Architect
- Experience in mapping AI use cases to intellectual property blocks
- Experience with hardware memory management units, prefetching, or memory subsystems
- Experience with multi-core systems and memory coherency
- Experience in machine learning model optimization on graphics processing unit architecture
Benefits
- Annual bonus eligibility
- Comprehensive compensation package with generous benefits
Skills and tech stack
Questions candidates ask
Where is this position located?
This position will be performed in an office setting located at 665 Clyde Avenue in Mountain View, California.
What is the salary range for the role?
The base pay range for roles at this level is $242,500 to $376,650 USD. Individual rates depend on factors including the role's function and location, as well as the individual's knowledge, skills, experience, education, and training.
What kind of products does the lab support?
The team explores innovative system-on-chip architecture for future Samsung Galaxy products, including smartphones, Internet of Things devices, and future hardware, echoing the company's artificial intelligence strategy.
What tools and models will I work with daily?
Daily responsibilities involve working with CNN and LLM models, programming in Python, SystemC, and C++, utilizing ML frameworks like TensorFlow, PyTorch, and ONNX, and performing modeling using Platform Architect.
The application is on the official Samsung Research America careers page.
Apply on company site