Sr. Software Development Engineer
About The Position
AMD is hiring a Sr. Software Development Engineer to research, design, develop, and test high-performance collective communication libraries, GPU communication runtimes, and benchmarking tools for AI/HPC workloads on AMD Instinct GPUs and next-generation accelerator platforms. The role develops, optimizes, troubleshoots, and debugs software for efficient multi-GPU and multi-node communication across AMD’s ROCm software stack. The engineer contributes to open-source GPU communication libraries, including RCCL, and benchmarking tools such as TransferBench. The work aims to maximize bandwidth utilization and minimize latency across PCIe, Infinity Fabric (xGMI), and high-speed network interconnects such as RDMA, RoCE, and InfiniBand. The position also develops and tunes collective algorithms for distributed training and inference frameworks, and creates micro-benchmarks and performance characterization tools for GPU-to-GPU, GPU-to-NIC, and host-to-device transfers. The role collaborates with hardware, firmware, driver, and AI framework teams to bring up new silicon, validate interconnect performance, and resolve cross-stack functional and performance issues. It addresses complex problems that require analysis of architectural, system-level, and network-level factors, applying knowledge of GPU architecture, parallel computing, networking, and modern AI/HPC workloads.
Requirements
- Master’s degree or foreign equivalent in Computer Science, Computer Engineering, Electrical Engineering, or a related field.
- Experience with Python, C, or C++.
- Experience developing benchmarks to test software performance.
- Experience in GPU software programming and development.
- Experience with software debugging on macOS, Linux, or Unix.
- Experience with data structures and multithreading.
- Experience with CUDA.
- Required skills and experience may be gained through graduate-level coursework, research, or internship experience.
Responsibilities
- Research, design, develop, and test high-performance collective communication libraries, GPU communication runtimes, and benchmarking tools for AI/HPC workloads.
- Develop, optimize, troubleshoot, and debug software enabling efficient multi-GPU and multi-node communication across AMD’s ROCm software stack.
- Contribute to open-source GPU communication libraries, including RCCL, and benchmarking tools including TransferBench.
- Develop and tune collective algorithms such as AllReduce, AllGather, ReduceScatter, Broadcast, and AllToAll for distributed training and inference frameworks.
- Design micro-benchmarks and performance characterization tools to evaluate GPU-to-GPU, GPU-to-NIC, and host-to-device transfer performance.
- Use performance results to influence hardware, firmware, and topology decisions for upcoming AMD Instinct products.
- Collaborate with hardware, firmware, driver, and AI framework teams to bring up new silicon, validate interconnect performance, and resolve cross-stack functional and performance issues.
- Analyze complex performance problems across architectural, system-level, and network-level factors.
Benefits
- AMD benefits are referenced as available at a glance; specific benefits are not listed.
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
