Senior Cloud & AI Workload Performance Engineer

Advanced Micro Devices; Town of Texas (WI); On-site; USD 140,000 - 210,000; Full time

Job summary

AMD is seeking a systems-minded performance engineer in Austin, TX to own workloads exposed to hyperscaler CPUs, AI systems and accelerator utilization. You will measure cloud-native performance, price-performance and performance-per-watt, comparing AMD EPYC against Intel and Arm platforms.

You will move from containers and orchestration to NUMA and memory bandwidth, explaining business impact with concise, evidence-based results. Collaboration with architecture teams is expected.

Qualifications

• Hands-on benchmarking performance of cloud, distributed or accelerated systems.
• Strong Linux systems skills and scripting for deployment/analysis.
• Experience with containers, orchestration and cloud-native workloads at scale.
• Ability to design fair experiments across different CPU architectures.
• Familiarity with hardware performance counters and system telemetry.

Responsibilities

• Define cloud-native and AI workload taxonomy for analysis and forecasting.
• Design reproducible methods for web services, containers, and scale-out data processing.
• Compare AMD/Intel/Arm/cloud environments with aligned software versions.
• Analyze host-CPU impact on accelerator utilization and end-to-end throughput.
• Model performance-per-watt, density and price-performance.
• Track factors affecting cross-ISA deployability.
• Explain bottlenecks with architecture experts.
• Forecast scaling trends and assess risks.
• Develop cross-ISA methodology documentation.
• Support validation with partners and produce technical reports.

Skills

Benchmarking

Linux systems

Containers & orchestration

AI workloads

CPU-accelerator data

Education

Bachelors or Masters in electrical or computer engineering

Tools

Automation scripts

Profiling tools

Performance counters

Job description

AMD is seeking a systems-minded performance engineer in Austin, TX to own workloads exposed to hyperscaler CPUs, AI systems and accelerator utilization. You will measure cloud-native performance, price-performance and performance-per-watt, comparing AMD EPYC against Intel and Arm platforms.

You will move from containers and orchestration to NUMA and memory bandwidth, explaining business impact with concise, evidence-based results. Collaboration with architecture teams is expected.