France's NCC is teaming up with AMD to host an online four-half-day training on scaling AI with AMD from 13 to 16 October 2025. This program offers a structured approach for AI researchers, data scientists, and practitioners to learn how to efficiently run and scale AI workloads on AMD GPUs. It features a mix of lectures, demos, and hands-on exercises using the AMD Accelerator Cloud, giving participants both theoretical knowledge and hands-on experience.
The program begins on 13 October 2025, with a beginner-friendly session introducing AMD GPU architecture, machine learning environments, and the AMD AI Assistant. The following three days focus on advanced topics, including HIP-Python, CuPy, MPI4Py, RCCL, PyTorch, AI for scientific applications, LLM-assisted optimization, and AMD profiling tools. This tiered approach makes the training accessible to participants with varying levels of expertise, while also offering in-depth sessions for advanced users.
Prerequisites differ by level: beginners are expected to know Python and the basics of a computing environment, while advanced participants should be familiar with Linux, AI on HPC systems, profiling and optimization techniques, and fundamental hardware concepts. The course is delivered fully online, with remote access to AMD infrastructures.
By the end of the program, attendees will have the skills to effectively run, scale, and optimize AI workloads on AMD GPUs, positioning them to apply state-of-the-art frameworks and tools in both academic and industrial settings. Registration is mandatory to secure participation.
