Senior GPU Performance and Power Benchmark Engineer
About The Position
AMD is committed to contributing to the future of the computing ecosystem, such as by participating in the development of industry standard performance and power benchmarks. This individual will join AMD’s Data Center Solutions Group (DSG) Architecture and Strategy team, a highly experienced organization of hardware and software engineers driving innovation across the full server stack—from node-level and rack-level design to security, cloud computing, memory technologies, high-performance fabrics, and beyond. In this role, the candidate will lead AMD’s involvement in shaping the next generation of industry standard GPU performance and power benchmarks. These benchmarks are expected to play a significant role in government energy-efficiency regulations worldwide and will help inform how the efficiency of GPU-based servers is evaluated, positioned, and communicated across the industry.
Requirements
- Self-starter that possesses excellent communication, leadership and interpersonal skills to be able to interact both inside and outside of AMD.
- Can overcome complex hurdles with minimal support, while keeping management appropriately informed and escalating important decisions when warranted.
- Be well-spoken with strong leadership, debate and influencing skills, especially in peer groups
- Have excellent problem-solving skills, an inquisitive mindset, and a strong focus on automation and technical investigation.
- Bachelor’s or higher degree in Electrical or Computer Engineering
Nice To Haves
- Strong knowledge of Artificial Intelligence (AI) and machine learning concepts, with practical experience applying these techniques to solve real-world problems through research or industry work.
- Deep technical understanding of server GPU hardware and software architecture, along with solid knowledge of server system and CPU architecture, including GPU/AI accelerator hardware-software interactions.
- Experience with GPU programming platforms like CUDA or HIP and deep learning frameworks, such as PyTorch, Triton, or JAX.
- Strong programming and scripting skills in languages such as Python, C++ and Java and using tools like Git.
- Understanding of multi-GPU communication technologies, such as UALink, NVLink, RCCL, and NCCL topologies and their performance characteristics.
- Advanced Linux OS and system expertise, including containers (e.g. Docker), system bring-up, debugging, and configuration, as well as familiarity with Windows.
- Experience with power and performance benchmarking, including experience with CPU performance counters, workload profiling and software profiling/debugging tools.
- Advanced data analysis skills, especially in support of performance characterization, benchmarking, and optimization.
- Experience with industry standards organization, especially benchmark development organizations, such as SPEC, MLCommons, or TPC, a strong plus.
Responsibilities
- Represent AMD externally to industry standard benchmark development organizations, including attending remote and in person meetings
- Lead AMD strategy for the development and measurement of a series of industry standard GPU performance and power benchmarks
- Measure performance and power on AMD and competitive GPU servers, including collection of relevant telemetry data, such as hardware performance counters.
- Investigate and root cause performance or power anomalies related to in development benchmarks, with support from relevant experts
- Develop and present benchmark design or adjustment proposals
- Internally coordinate with and inform GPU software and hardware architecture, product management and marketing teams
Benefits
- AMD benefits at a glance.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
