PhD HPC AI Research Intern - GPU Performance

AMD; Santa Clara (CA); Hybride; USD 62 000 - 82 000; Plein temps

Résumé du poste

AMD invites PhD students to join our research team at the intersection of High-Performance Computing and Artificial Intelligence. You will optimize AI workloads for large-scale computing environments and contribute to algorithms for future AMD GPUs, with hybrid or onsite work in San Jose or Santa Clara.

You will build and analyze benchmarks, assess development tools and runtimes, and explore code optimization techniques.

Qualifications

• Enrolled in a PhD program in Computer Science, Computational Science, Electrical/Computer Engineering, Applied Mathematics, or a related field.
• Strong background in parallel computing, distributed systems, or AI/ML frameworks.
• Proficiency in programming languages such as Python, C/C++, Fortran or ROCm/CUDA.
• Experience with at least one deep learning framework (e.g., PyTorch, TensorFlow, JAX).
• Familiarity with MPI, OpenMP, or GPU programming.
• Solid understanding of numerical methods, optimization, or scientific computing.

Responsabilités

• Build, run, and analyze performance of benchmarks on GPU accelerated platforms.
• Assess the capabilities of development tools and runtime environment in terms of capabilities, performance, and usability.
• Explore the benefits of different code optimization techniques

Connaissances

PhD student

Parallel computing

AI/ML frameworks

Python

C/C++

Formation

PhD program in CS/Computational Science/EC/ECE/Applied Math

Outils

MPI

OpenMP

GPU programming

ROCm/CUDA

Description du poste

AMD invites PhD students to join our research team at the intersection of High-Performance Computing and Artificial Intelligence. You will optimize AI workloads for large-scale computing environments and contribute to algorithms for future AMD GPUs, with hybrid or onsite work in San Jose or Santa Clara.

You will build and analyze benchmarks, assess development tools and runtimes, and explore code optimization techniques.