Samsung
Staff Engineer, GPU Memory Architect
$151K–$252K
9808 Scranton Rd, San Diego, CA, USA +1 · On-site · Staff · Full-time · Posted Oct 9, 2026
- Location
- 9808 Scranton Rd, San Diego, CA, USA
3655 N 1st St, San Jose, CA, USA - Experience
- 5+ years with a Bachelor’s Degree; or 3+ years with a Master’s Degree; or 1+ years with a Ph.D. Also requires 5+ years in GPU architecture, design, or development.
- Education
- Bachelor’s Degree in Computer Science/Engineering, or Master’s Degree, or Ph.D.
- Visa sponsorship
- Applicants must be authorized to access information subject to U.S. export control laws or eligible for government authorization.
About the job
Architect and develop GPU memory subsystems and on-chip interconnects for next-generation Samsung Exynos GPUs, optimizing performance, power, and area for graphics and AI workloads.
What you'll do
- Lead architecture studies and microarchitecture development for memory hierarchies, caches, and high-bandwidth memory interfaces.
- Develop PPA models and run trade studies for workloads including ray tracing, AI inference, and large-scale graphics.
- Coordinate integration with cross-functional teams and produce roadmaps, design specifications, and validation plans.
What they ask for
- 5+ years of GPU architecture, design, or development experience focused on high-performance, low-power designs.
- Knowledge of machine learning or graphics and GPU pipelines, hardware design, and computer architecture.
- Experience analyzing and optimizing memory subsystems, with knowledge of memory hierarchies, bandwidth allocation, coherence protocols, and on-chip interconnect.
- Strong background in low-power design techniques and performance modeling tools, such as analytical models and RTL-level simulation.
Nice to have
- Experience with CPU, NPU, or other complex processing units, such as ARM, x86, or RISC-V.
- Experience in performance modeling, statistical analysis, and hardware-software co-design for AI accelerators.
Pay
$151,000 and $251,800 (base pay range; the posting does not specify an ISO currency code).
Tools
ARM, x86, RISC-V, SRAM, DRAM, RTL-level simulation
AI focus
Architect GPU memory systems for emerging AI workloads, including AI inference; translate algorithmic requirements into memory-system specifications.
Model-written summary · the original posting is authoritative