AI Framework and Performance Experts(vLLM)
Core
Building and optimizing production-grade single-node and distributed inference runtimes for large language models on AMD GPUs.
Role type
Senior IC systems-minded ML engineer (LLM inference frameworks)
Builds
Production inference runtimes, benchmark-grade pipelines, and upstream contributions to open-source frameworks (vLLM, SGLang)
Domain
AI/ML Infrastructure, High-Performance Computing, Distributed Systems
Deliverable
production ML models
Required skills
Distributed LLM inference runtime architecture, TP/PP/EP (MoE) execution design, KV-cache management, multi-node pipeline implementation (RCCL, RDMA), GPU backend integration (ROCm, HIPBLAS-LT), Compiler optimization (Triton, LLVM), Python, C/C++, High-throughput inference optimization
Preferred skills
vLLM/SGLang stack expertise, Open-source contribution track record, PyTorch/TensorFlow integration, NVIDIA/AMD GPU architecture knowledge
Technologies
vLLM, SGLang, llm-d, ROCm, RCCL, RDMA, Triton, LLVM, HIPBLAS-LT, FlashAttention, MLA
Responsibilities
Architect and optimize distributed LLM inference runtimes; Design hybrid execution strategies (TP/PP/EP); Implement multi-node inference pipelines; Drive throughput, latency, and memory efficiency; Collaborate with compiler and kernel teams; Upstream features to open-source frameworks; Enable customer PoCs and production deployments
Seniority
Senior, hands-on IC
