Overview:
WHAT YOU DO AT AMD CHANGES EVERYTHING
At AMD, our mission is to build great products that accelerate next-generation computing experiences—from AI and data centers, to PCs, gaming and embedded systems. Grounded in a culture of innovation and collaboration, we believe real progress comes from bold ideas, human ingenuity and a shared passion to create something extraordinary. When you join AMD, you’ll discover the real differentiator is our culture. We push the limits of innovation to solve the world’s most important challenges—striving for execution excellence, while being direct, humble, collaborative, and inclusive of diverse perspectives. Join us as we shape the future of AI and beyond. Together, we advance your career.
Responsibilities:
Software Engineer — High-Performance GPU Communication (MORI)
Role: Software Engineer / Systems Engineer, GPU Networking & Inference Infrastructure Team: ROCm System Software — Communication Primitives
About the Project
MORI (Modular RDMA Interface) is an open-source framework powering AMD GPU communication in large-scale LLM inference. It provides the RDMA + GPU-direct networking layer for MoE expert parallelism and prefill/decode disaggregation in SGLang and vLLM, and owns the KVCache management and storage layer via MORI-UMBP (Unified Memory & Bandwidth Pool). MORI achieved state-of-the-art results in the SemiAnalysis InferenceX v2 evaluation.
What You'll Work On
Inference framework integration: Own end-to-end integration of MORI primitives into SGLang and vLLM — Python operator APIs, MORI-EP dispatch/combine in MoE forward passes, and MORI-IO in KVCache transfer pipelines.
MORI-UMBP: Integrate tiered KVCache storage and distributed key-value access into inference serving stacks.
PD disaggregation: Integrate MORI-IO into the prefill/decode path, enabling high-throughput KVCache transfers over GPU-direct RDMA.
Expert Parallelism (EP): Land and maintain MORI-EP in SGLang and vLLM, covering scheduling, routing, and EPLB for MoE models like DeepSeek V3 across 8–64 GPUs.
MORI-SHMEM: Integrate and maintain the symmetric memory runtime that underpins all MORI components — managing symmetric GPU memory allocation, RDMA transport initialization (IB, AINIC, Thor2), P2P/XGMI address translation, and device-side state for GPU kernels via MORI-IR bitcode.
Performance benchmarking: Design and run end-to-end benchmarks (throughput, TTFT, ITL) across EP and PD disagg configurations; drive optimization from profiling data.
Qualifications
Required: - Deep familiarity with at least one major LLM inference framework (SGLang, vLLM, TensorRT-LLM, or equivalent) — scheduler, attention backend, KVCache manager, and distributed execution engine. - Strong understanding of LLM serving: MoE expert parallelism, prefill/decode disaggregation, KVCache reuse, tensor/pipeline/sequence parallelism. - Solid C++ and Python; comfortable in a mixed C++/HIP/Python codebase with PyTorch custom operator extensions. - Experience contributing to large open-source projects: upstream PRs, code review, cross-team coordination.
Nice to Have: - RDMA concepts: verbs API, queue pairs, completion queues, memory registration, GPUDirect Async (IBGDA). - Collective communication libraries (NCCL, RCCL, MPI) and their integration into distributed stacks. - GPU cluster network topologies: XGMI/NVLink (intra-node), InfiniBand/RoCE (inter-node), and their impact on MoE all-to-all patterns. - NIC vendor ecosystems (Mellanox ConnectX, AMD Pollara/AINIC, Broadcom Thor2) and userspace driver libraries. - Profiling network-bound workloads with rocprofv3, Perfetto, ibstat/perfquery. - ROCm, hipcc, or AMD GPU architecture experience.
What Makes This Role Unique
You'll own the bridge between MORI's low-level GPU networking layer and inference frameworks running trillion-parameter models at scale. MORI-EP and MORI-IO are already merged into SGLang and vLLM — your contributions ship directly to open-source and production. The team brings deep RDMA and GPU kernel expertise, so you can focus on inference-level impact while learning the networking layer from those who built it.
#LI-FL1
Qualifications:
Benefits offered are described: AMD benefits at a glance.
AMD does not accept unsolicited resumes from headhunters, recruitment agencies, or fee-based recruitment services. AMD and its subsidiaries are equal opportunity, inclusive employers and will consider all applicants without regard to age, ancestry, color, marital status, medical condition, mental or physical disability, national origin, race, religion, political and/or third-party affiliation, sex, pregnancy, sexual orientation, gender identity, military or veteran status, or any other characteristic protected by law. We encourage applications from all qualified candidates and will accommodate applicants’ needs under the respective laws throughout all stages of the recruitment and selection process.
AMD may use Artificial Intelligence to help screen, assess or select applicants for this position. AMD’s “Responsible AI Policy” is available here.
This posting is for an existing vacancy.
