Principal Distributed Systems Software Engineer, Collectives
Principal engineer on AMD's Collectives and Network Performance team, defining strategy and optimizing collective communication libraries (RCCL/NCCL-style) and scale-up/scale-out networking software for distributed AI training and inference on AMD GPUs. Deep C++/Linux performance work spanning ROCm, NICs, PyTorch, JAX, vLLM and SGLang. Onsite in San Jose or Austin; no visa sponsorship.
- Compensation
- Not disclosed
- Location
- US,CA,San Jose
- Work setup
- Not specified
- Seniority
- Principal
- Employment
- Not specified
- Experience
- Not specified
- Category
- Software engineering
- Posted
Skills
Must have
- Industry-leading expert in collective communication and network performance
- Advanced technical and analytical skills in Linux environments
- Deep C++ expertise
- Technical leadership: guiding teams and driving projects and initiatives independently
- Deep knowledge of network, NIC and GPU hardware architecture
- Hands-on experience mapping model architectures to low-level software and hardware
- Strong grasp of current trends in distributed training and inference at scale
- Master's and/or PhD in Computer Science, Computer Engineering, Electrical Engineering, or related field
Nice to have
- Deep expertise designing/optimizing collective communication libraries (RCCL, NCCL, MPI)
- Thorough understanding of NIC, RDMA, scale-up and scale-out network architectures and their impact on model performance
- Proven experience in performance analysis, projection and bottleneck identification across the full stack, from model to hardware
- Ability to integrate optimized communication into ML frameworks such as PyTorch, JAX, vLLM and SGLang
- Advanced proficiency in C++ and Python with performance tuning, debugging and robust test design
- Broad, in-depth experience with large-scale heterogeneous compute environments and cluster-level AI workload optimization
- Demonstrated ability to lead cross-team initiatives, mentor engineers, and communicate effectively in writing and presentations
- Senior-level industry experience in technical software development focused on performance engineering, communication libraries, networking, and GPU/distributed systems
Full description
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 a core member of the Collectives and Network Performance team, you will play a pivotal role in defining and delivering industry-leading AI pre-training and distributed inference performance on AMD GPUs. Your expertise will be critical in optimizing collective communication libraries, scale-up and scale-out networking software, and training/inference performance across multi-GPU and multi-node systems. You will engage with internal hardware, GPU library, compiler, runtime and ROCm teams as well as open-source maintainers to ensure seamless integration of optimizations, applying advanced engineering principles to drive continuous improvement across multiple generations of AMD GPUs and a wide range of state-of-the-art AI models.
THE PERSON:
Seeking an industry-leading expert in collective communication and network performance, with advanced technical and analytical skills in Linux environments and deep C++ expertise. The ideal candidate will excel in providing technical leadership, guiding teams, and driving projects and initiatives independently. You will define goals, scope, and own development efforts while collaborating effectively within a high-performing team. You have deep knowledge of network, NIC and GPU hardware architecture, hands-on experience mapping model architectures to low-level software and hardware, and a strong grasp of current trends in distributed training and inference at scale.
KEY RESPONSIBILITIES:
- Drive Collectives & Network Strategy: Define goals, scope and roadmap for a major area of AMD’s collectives and network optimizations, and align them with hardware, framework and ROCm roadmaps.
- Architect Collective Communication Libraries: Lead the design and development of collective algorithms and optimizations (e.g., all-reduce, all-to-all, all-gather) across scale-up and scale-out networks.
- Optimize Distributed Training & Inference: Lead performance tuning, profiling and analysis of large-scale LLM, diffusion, multimodal, RecSys and generative AI models on single-node and multi-node systems, and drive the tradeoffs and design decisions.
- Hardware-Software Co-Design: Participate in co-design of future scale-up networks, NICs and scale-out networks, and represent software performance requirements to hardware teams.