Principal AI Performance and Tools

What you'll do

• Lead the technical strategy for AI performance analysis, benchmarking, profiling, and optimization across AMD platforms.
• Translate customer requirements and performance gaps into prioritized engineering features and software improvements.
• Analyze inference and training workloads to identify bottlenecks and recommend architectural optimizations.
• Develop rigorous performance narratives including benchmark methodologies, workload characterization, and competitive analysis.
• Influence roadmaps for compilers, runtimes, and performance-tuning tools across the AI software ecosystem.
• Define reference workflows and best practices for optimizing AI models and applications.
• Engage with strategic customers and partners to understand real-world development workflows and tooling needs.
• Represent AMD in technical forums, industry events, and open-source communities to communicate technical capabilities.

What we're looking for

• Bachelor's or Master's degree in Computer Science, Computer Engineering, Electrical Engineering, Business, or a related field.
• Proven years of experience in software engineering, systems architecture, performance engineering, AI/ML infrastructure, or developer tools (preferred).
• Hands-on expertise with AI and machine learning frameworks such as PyTorch, TensorFlow, JAX, ONNX Runtime, vLLM, or similar technologies (preferred).
• Strong understanding of GPU computing, heterogeneous systems, model serving, and AI optimization (preferred).
• Experience with performance profiling, benchmarking, tracing, debugging, compiler technologies, runtimes, or systems observability (preferred).
• Proficiency in one or more programming languages including Python, C++, C, HIP, CUDA, or similar languages (preferred).
• Experience optimizing large language models, generative AI, recommendation systems, computer vision, or scientific computing workloads (preferred).
• Familiarity with cloud AI platforms, Kubernetes-based deployments, model-serving stacks, and enterprise AI operations (preferred).