Staff Engineer, GPU Memory Architect
Core
Define, architect, and implement next-generation GPU memory subsystems and on-chip interconnects for Samsung's Exynos chipset to optimize performance, power, and area (PPA) for mobile and AI workloads.
Role type
Staff GPU Memory Architect (IC)
Builds
GPU global memory hierarchies, cache structures, high-bandwidth memory interfaces, and on-chip interconnects for SoC integration.
Domain
Semiconductor / Hardware Architecture / GPU
Deliverable
production ML models | product features
Required skills
GPU architecture design, memory hierarchy optimization (caches, SRAM, DRAM), on-chip interconnect protocols, performance modeling, low-power design techniques, architectural trade-studies, SoC integration, hardware/software co-design for AI/ML.
Preferred skills
CPU/NPU architecture (ARM, x86, RISC-V), statistical analysis, hardware-software co-design for AI accelerators.
Technologies
Exynos chipset, RTL simulation, analytical modeling tools, ray-tracing pipelines, AI inference workloads.
Responsibilities
Lead architecture studies and micro-architecture development for GPU memory; define innovative features for memory subsystems; collaborate with cross-functional teams for SoC integration; build and maintain PPA models; produce architectural roadmaps and design specifications.
Seniority
Staff, hands-on IC with strategic influence
