Intern, Ai/Llm Characterization & Tight Coupling Analysis

What You'll LearnProject: Supercomputing Project (SCP), Modeling Track: Super node hardware sub-team & AI/LLM/GPT Application sub-teamSkills You'll Learn:Understanding how various coupling architectures can impact application performance and power.Running of applications using various profiling tools on various hardware.Analyzing profiling results to understand application and hardware bottlenecks and profiler limitations.Gain understanding of key phases of applications under study and their performance on various hardware.Running of applications on the SST simulator, understanding its configurations, and possibly the internal modeling details.What You'll DoCharacterize the amount of latency imposed by kernel launches and data transfers for various GPU-accelerated workloads on various hardware (discrete and integrated GPUs) using existing profiling tools (nsys/nsight, Vtune, Advisor, Tau).AI inference: LLMs, DLRMs, Graph NNsEstimate potential speedup of these workloads if run on tightly-coupled accelerator hardware. Explain assumptions/rationale for the estimates.Estimate potential power savings of these workloads if run on tightly-coupled accelerator hardware. Explain assumptions/rationale for the estimates.Identify specific AI workloads and specific areas/phases/kernels of AI workloads that benefit from tight coupling.As the research progresses, alternate or additional scope may be proposed by either party. Inclusion of that scope must be mutually agreed upon between Samsung and internship candidate.Complete other responsibilities as assigned.Location: Hybrid, working onsite at our San Jose, CA headquarters 3 days per week, with the flexibility to work remotely the remainder of your time  What You BringMasters or PhD candidate in Computer Science, Electrical Engineering, or Computer EngineeringExperience with HPC systems preferred In-Depth familiarity with high performance computer architecture (processors, network, or memory) preferred Familiarity with HPC and AI applications (e.g. LLM/GPT) preferredAn avid learner, you approach challenges with curiosity and resilience, seeking data to help build understanding.You're collaborative, building relationships, humbly offering support and openly welcoming approaches.Innovative and creative, you proactively explore new ideas and adapt quickly to change.