Intel GPU AI Compute Architect Intern 

India, Karnataka, Bengaluru 

567214737

We are an AI architecture team that leverages insights from AI workloads to inform future Intel hardware design decisions. Our work involves:1. Studying emerging AI workloads, including LLM and multimodality GAI2. Developing an internal performance analysis tool (e.g., ArchBench), including adding hardware architecture features and support for new AI workloads.3. Generating and analyzing performance data to inform architecture decisions, including calibrating performance projections with real silicon or simulators, and producing high-quality performance projections.We are seeking highly motivated candidates who are eager to learn about AI hardware, possess strong collaboration skills, and can adapt to a fast-paced software development environment. As an intern, you may be responsible for tasks such as:1. Generating and analyzing performance data of emerging AI workloads, such as multimodality generative AI.2. Co-developing a module of an internal performance analysis tool or improving its accuracy.3. Correlating simulation results with silicon measurements.4. Evaluating GPU AI hardware architecture.5. Providing guidelines for future software optimization, such as cache tiling.Minimum Qualifications:BS/MS in EECS, Mathematics, or related majors.Proficient in Python.Good data analysis and presentation skills.Preferred Qualifications:Knowledge of AI and deep learning, such as Large Language Models (LLM), Stable Diffusion, or multimodality generative AI.Experience with performance analysis tools/simulators.Experience with workload characterization.Experience with performance projection.Strong knowledge of computer architecture.Familiarity with PyTorch.