Senior System Performance Telemetry Architect

Arm Limited; Austin (TX); On-site; USD 250,000 - 338,000; Full time

Job summary

Arm Limited seeks a Principal System Performance Telemetry Architect to shape performance telemetry across CPUs, AI accelerators, GPUs, memory systems, and interconnects.

You will work with silicon and software teams to embed profiling capabilities, define data collection methods, and drive platform-wide observability for power-performance optimization and real-world deployment debugging.

Qualifications

• Proven experience in system architecture and performance engineering.
• Track record in profiling, workload characterization and bottleneck analysis.
• Ability to translate telemetry into architectural recommendations.

Responsibilities

• Define the performance telemetry and profiling architecture for CPUs, AI accelerators, GPUs, and memory systems.
• Drive platform specifications and reference architectures for observability across Arm-based infrastructure.
• Collaborate with SoC/IP/platform architects and software teams to embed profiling capabilities.
• Define telemetry requirements supporting performance modeling and power-performance optimization.
• Develop platform-level analysis strategies for heterogeneous systems and multi-die architectures.
• Ensure data collection and correlation from CPUs, caches, memory, interconnects, and accelerators.

Skills

Performance optimization

Profiling methodologies

System architecture

Cross-functional leadership

Hardware–software co-design

Communication leadership

Heterogeneous computing

Tools

NVIDIA Nsight

Intel VTune

Linux perf

Windows Performance Analyzer

Arm Streamline

Arm Performance Studio

Job description

Arm Limited seeks a Principal System Performance Telemetry Architect to shape performance telemetry across CPUs, AI accelerators, GPUs, memory systems, and interconnects.

You will work with silicon and software teams to embed profiling capabilities, define data collection methods, and drive platform-wide observability for power-performance optimization and real-world deployment debugging.