Memory Systems Performance & Workload Architect - DMTS/SMTS
Core
Lead workload-driven memory architecture and performance strategy across mobile and server-class platforms, translating real-world application behavior into next-generation memory architecture and product features.
Role type
Principal Memory Systems Performance & Workload Architect
Builds
Next-generation DRAM, LPDDR, and HBM based memory systems for servers and mobiles
Domain
Memory Systems Architecture, Workload Analysis, Performance Engineering for servers and mobiles
Deliverable
production ML models | product features
Required skills
Workload characterization at scale, Memory bandwidth/latency/QoS analysis, System-level tradeoff analysis, Performance modeling and simulation, Cross-stack optimization (OS/Firmware/Hardware), Linux kernel memory management, Benchmarking suite definition
Preferred skills
AI/ML workload optimization, Database and cloud-native systems expertise, Performance modeling frameworks, ARM and x86 system architectures
Technologies
Python, C/C++, Java, Linux, eBPF, PMU, SPEC, MLPerf, STREAM, RocksDB, MySQL, PostgreSQL, Redis, Memcached
Responsibilities
Define industry-leading benchmarking frameworks for mobile and server workloads, Characterize memory behavior (bandwidth, latency, access patterns, NUMA), Lead root cause analysis using PMU counters and perf tools, Drive system-level tradeoff analysis for product decisions, Establish performance modeling and trace-driven workload replay frameworks, Drive end-to-end optimizations across application to hardware stack
Seniority
Principal, strategy & mentorship
