Staff Engineer, SSD Storage and Systems Architecture

Samsung Semiconductor•San Jose, CA

9d•$163,000 - $253,000•Hybrid

About The Position

Samsung Semiconductor is seeking an industry-leading SSD Storage and System Architect with recognized technical influence to design and define our next-generation Enterprise, Datacenter (DC), and AI-optimized storage architectures. As AI/ML workloads explode, storage has become the primary bottleneck for data ingestion, checkpointing, and GPU utilization. This role focuses on the end-to-end design of high-performance storage systems, optimization of flash translation layers (FTL), and hardware-software co-design engineered specifically to accelerate AI training and inference. You will leverage your deep domain expertise to align system architecture with major hyperscale cloud vendor AI clusters while directly bridging the gap with Samsung's internal R&D teams to maximize the competitive edge, silicon efficiency, and performance synergy of Samsung's cutting-edge memory, CXL, and controller solutions.

Requirements

Bachelor’s degree in Electrical Engineering, Computer Engineering, Computer Science, or a related technical field with 10+ years experience, or a Master's degree with 8+ years experience, or a PhD with 5+ years experience.
Minimum of 10+ years of direct experience in storage system architecture, including foundational hands-on experience with legacy HDD system architecture and caching algorithms alongside modern NAND flash, SSD architecture, and AI/ML data pipeline storage.
Verified track record of architectural leadership, technical contributions to open-source storage/AI projects, or technical influence within the broader storage ecosystem.
Proven experience architecting or deploying storage systems that satisfy OCP Datacenter NVMe SSD Specifications and scale seamlessly within high-performance computing (HPC) or GPU-accelerated hyperscale datacenter environments.
Deep technical knowledge of PCIe/NVMe protocols, specialized enterprise features like GPUDirect Storage (GDS), FDP, ZNS, SSD controller internal architectures, and enterprise interfaces (NVMe-oF, CXL, SAS/SATA/SCSI).
Ability to quickly master and leverage Samsung's proprietary controller capabilities, vertical NAND (V-NAND) characteristics, CXL innovations, and DRAM caching to maximize architectural efficiency for AI workloads.
Exceptional written and verbal English communication skills, with a proven ability to pitch complex architectural designs, author technical specs, and build consensus across multi-disciplinary global teams.
Ability to evaluate storage as part of the complete system architecture, considering compute, memory, interconnect, networking, local storage, and shared storage together.
Ability to translate real application and AI workload behavior into clear storage architecture and SSD product requirements.
Ability to identify end-to-end system bottlenecks and determine whether performance limitations originate from compute, memory, PCIe fabric, software stack, or storage.
Ability to define optimal data placement and tiering across memory, cache, local NVMe, shared storage, and capacity tiers based on workload characteristics.
Ability to balance performance, latency, capacity, endurance, power, thermal, reliability, scalability, and cost at the system level.
Ability to identify emerging storage architecture needs and propose new system and SSD concepts for next-generation AI and data center workloads.
Ability to analyze complex issues across application, system software, platform hardware, and storage boundaries and drive architectural solutions rather than point fixes.
Ability to establish architectural direction, challenge assumptions, evaluate alternatives, and drive technical decisions across engineering and product teams.

Responsibilities

Define and model enterprise/DC SSD architectures optimized for massive parallel read paths required by AI data loaders (e.g., PyTorch, TensorFlow) and high-burst write paths for AI model checkpointing.
Translate complex hardware, CXL, and controller constraints into optimized firmware architectures for Samsung’s enterprise SSD controller, FTL, and NAND flash subsystems.
Architect features tailored for hyperscaler AI infrastructure, matching specifications like Microsoft’s Azure AI Cloud storage architectures and Meta’s next-generation GPU-accelerated storage platforms.
Collaborate closely with ASIC design, firmware development, and validation teams to resolve cross-domain architectural bottlenecks in ultra-low latency environments.
Develop architectural simulators or FPGA prototypes to evaluate next-generation storage features, AI caching algorithms, and host-managed storage paradigms.
Leverage an established technical network to align emerging system architectures with broader storage industry trends, open-source AI ecosystems, and storage standards (NVMe, OCP, SNIA).
Drive the creation of foundational architectural patents in the domains of AI storage optimization and near-data processing to protect Samsung's system-level intellectual property.
Jointly architect next-generation AI datacenter deployment models (e.g., custom pooling, disaggregated storage, GPUDirect Storage [GDS] integrations, and Ethernet-attached SSDs) directly with lead architects from AWS, Google Cloud, Microsoft, and Meta.
Lead the architectural definition of advanced storage paradigms such as FDP (Flexible Data Placement), ZNS (Zoned Namespaces), and computational storage specifically tailored to offload AI data-preprocessing tasks.