Sr. Distributed Systems Software Engineer, Collectives

Position: Sr. Staff Distributed Systems Software Engineer, Collectives
ADVANCE YOUR CAREER. ADVANCE THE WORLD. At AMD, we believe technology has the power to solve the world’s most important challenges. From advancing healthcare and scientific discovery to powering AI and the technologies people rely on every day, innovation at AMD is shaping the future. Whether you’re designing next-gen processors, enabling AI breakthroughs, or bringing leading edge products to market, every role at AMD contributes to something bigger — technology that moves the world forward.

Join us and, together, we’ll advance your career.

THE ROLE:

As an experienced member of the Collectives and Network Performance team, you will play an important role in optimizing collective communication and networking software for AMD GPUs. You will enhance communication libraries and training/inference performance across multi-GPU and multi-node systems, working with internal GPU library, runtime and ROCm teams and with open-source maintainers to integrate optimizations across multiple generations of AMD GPUs and a range of state-of-the-art AI models.

THE PERSON:
Seeking an experienced C++ developer with strong technical and analytical skills in Linux environments. The ideal candidate can take ownership of features and projects with limited guidance, contribute to defining scope and technical approach, and collaborate effectively within a high-performing team. You have solid knowledge of network, NIC and GPU hardware architecture, hands-on experience profiling and optimizing distributed AI workloads, and a good understanding of modern training and inference optimization.

You also support the growth of less experienced engineers.

KEY RESPONSIBILITIES:

Develop Collective Communication Features:
Design, implement and optimize features in collective communication libraries across scale-up and scale-out networks.

Optimize Distributed Training & Inference:
Profile, tune and analyze LLM, diffusion, multimodal, Rec Sys and generative AI workloads on single-node and multi-node systems, and recommend design tradeoffs.

Own Features End to End:
Drive assigned features and projects from design through delivery with limited guidance, including scoping, implementation, testing and performance validation.

Performance Modeling & Tooling:
Develop and improve tools and infrastructure for benchmarking, performance estimation, regression tracking and reporting.

Collaborate with Framework & Library Teams:
Work with internal GPU library, runtime and ROCm teams, and with open-source maintainers, to ensure code changes meet requirements and are integrated upstream.

Hardware-Software Co-Design Input:
Contribute software performance data and analysis to discussions on future scale-up networks, NICs and scale-out networks.

Customer & Stakeholder Support:
Help customers and internal teams apply effective load balancing, workload scheduling and model sharding strategies, and present results and recommendations.

Software Engineering Best Practices:
Apply sound engineering principles to deliver robust, maintainable solutions.

Guide & Mentor:
Mentor junior engineers through code reviews, knowledge sharing and technical guidance.

PREFERRED EXPERIENCE:

Collective Communication & Network Knowledge:
Strong knowledge of collective communication libraries (e.g., RCCL, NCCL, MPI) and a good understanding of NIC, RDMA and scale-up and scale-out network architectures.

Performance Analysis:
Strong experience in profiling, performance analysis and bottleneck identification for distributed AI workloads.

Deep Learning Integration:
Hands-on experience integrating optimized compute or communication into ML frameworks such as PyTorch, JAX, vLLM or SGLang, with a focus on throughput and scalability.

Software Engineering Excellence:
Strong proficiency in C++…