Python, analytical performance models, AI workload expertise, NPU AI-accelerator architecture, GPU architecture, execution-driven architectural cycle-accurate simulators

Python, C++, Pcie, analytical performance models, GPU architecture, kernel execution graphs, Linux tracing, LLM and Transformer performance, simd, SIMT models, cache hierarchy, tensor pipeline, custom fabrics, data parallelism, architectural cycle-accurate simulators, high-bandwidth interconnects, memory technologies, performance counters