Intel is a global technology leader that designs and manufactures semiconductor chips, cloud‑computing solutions, and AI‑driven platforms. With a strong focus on innovation, Intel invests heavily in research labs across the world, including India, to push the boundaries of compute performance, security, and sustainability. The company’s culture blends cutting‑edge engineering with a collaborative environment, offering employees opportunities to work on products that power everything from data‑centers to personal devices. The AI System Performance Engineer role in Bengaluru is tailored for recent graduates or early‑career engineers who are passionate about the intersection of artificial intelligence and hardware architecture. The incumbent will be responsible for building analytical models that predict how AI training and inference workloads behave on current and future GPU designs, evaluating multi‑GPU communication patterns, and identifying bottlenecks that affect throughput and latency. This position provides exposure to Intel’s internal simulation frameworks, open‑source performance tools, and real‑world AI workloads such as large language models. Key Responsibilities: 1. Develop and maintain analytical models for AI training and inference performance across various GPU configurations. 2. Simulate and evaluate memory hierarchy, interconnect bandwidth, and compute utilization for large‑scale AI workloads. 3. Conduct comparative studies of single‑GPU versus multi‑GPU deployments, focusing on communication overhead and scaling efficiency. 4. Adapt open‑source simulators to accurately reflect Intel‑specific architectural features. 5. Generate performance reports and recommendation documents for hardware and system architects. 6. Collaborate with software teams to integrate performance insights into compiler and runtime optimizations. 7. Stay updated on emerging AI workloads, especially LLM training and inference trends, and incorporate findings into models. 8. Participate in cross‑functional design reviews, providing data‑driven feedback on architectural decisions. 9. Automate data collection and analysis pipelines using Python and related libraries. 10. Mentor interns or junior team members on performance modelling techniques. Tech Stack: Python, NumPy/Pandas, performance simulation tools (e.g., gem5, Accel‑Sim), GPU architecture fundamentals, Linux environment, Git, basic knowledge of CUDA/HIP. Growth Path: Starting as an Engineer I, the role can progress to Senior Engineer, Principal Engineer, and eventually to Architecture Lead or AI Systems Manager, with opportunities to move into product management or research. Why Join Intel: Intel offers a world‑class learning ecosystem, mentorship from industry veterans, and the chance to influence next‑generation AI hardware. The hybrid work model ensures flexibility while maintaining strong team cohesion, and the company’s commitment to diversity and inclusion creates a supportive environment for fresh talent.