Proficiency in Parallel Paradigms and Profiling Tools: Advanced knowledge of parallel programming models (MPI, OpenMP, CUDA, ROCm) paired with hands-on expertise in using performance analysis, profiling, and tracing tools (e.g., NVIDIA Nsight, Intel VTune, Score-P) to pinpoint bottlenecks in compute, memory, and I/O. Essential Requirements: Deep System and Architecture Expertise: A comprehensive understanding of heterogeneous HPC architectures-including modern CPUs, GPUs, high-speed fabrics (InfiniBand), memory hierarchies, and parallel file systems (BeeGFS, Pixstor)-along with the ability to tune BIOS, OS, and system-level parameters for maximum throughput.