Principal GenAI Inference Optimization Engineer
Core
Optimize performance, efficiency, and scalability of generative AI inference workloads on AMD GPU platforms for real-world deployment of large-scale models.
Role type
Principal GenAI Inference Optimization Engineer (IC)
Builds
Production inference serving systems for LLM and multimodal models
Domain
AI/ML inference, GPU systems, high-performance computing
Deliverable
production ML models
Required skills
GenAI inference optimization, GPU architecture, distributed systems, profiling tools, C++/CUDA/HIP, Python, quantization, batching strategies, speculative decoding
Preferred skills
Large-scale serving systems, kernel optimization, memory hierarchy, communication patterns, ML frameworks (PyTorch/JAX/TensorFlow), open-source contributions
Technologies
vLLM, SGLang, Triton, TensorRT-LLM, PyTorch, JAX, TensorFlow, CUDA, HIP, PCIe, Infinity Fabric, RDMA
Responsibilities
Optimize GenAI inference workloads across single-node and distributed environments; Analyze and resolve bottlenecks in compute, memory, and communication; Implement optimization techniques like quantization and prefix caching; Develop profiling and benchmarking tools; Collaborate with hardware and compiler teams; Document performance guidelines
Seniority
Principal, hands-on IC
