AMD Launches ROCm Certification to Build AI, HPC Skills

AMD has announced the launch of its ROCm Certification Program, aimed at equipping developers and engineers with hands-on skills for AI and high-performance computing (HPC) on AMD Instinct GPUs. The first certification level, Associate, will be available starting July 24, 2026.

This initiative comes as demand surges for GPU-based solutions in AI and HPC. Organizations increasingly rely on multi-GPU and multi-node clusters, and validated expertise in AMD’s open ROCm software stack is becoming a sought-after credential. ROCm (Radeon Open Compute) powers workloads on AMD Instinct GPUs, such as the MI300 series, which AMD has positioned as a cornerstone for data center AI deployments.

Certification Levels and Their Focus

The ROCm Certification Program is structured into three levels:

• Associate (launching July 24): Designed for developers new to AMD GPUs or transitioning from other platforms, this level emphasizes ROCm fundamentals, HIP programming, and optimizing single-GPU AI and HPC workloads.
• Professional (launching late 2026): Targets multi-GPU deployments and distributed AI training, covering tools like RCCL and frameworks like vLLM for inference scaling.
• Expert (early 2027): Focuses on end-to-end production systems, including Kubernetes-based MLOps pipelines and fleet-scale GPU infrastructure management.

Each certification requires hands-on labs and practical assessments, ensuring participants can translate theoretical knowledge into real-world applications. This practical focus differentiates the ROCm Certification from other programs in the AI and HPC space.

Why It Matters Now

AMD’s push for ROCm adoption aligns with its broader GPU strategy. The company has been aggressively expanding its ecosystem, as evidenced by recent updates like ROCm 7.2.4 (released May 29, 2026), which improved AI inference performance, and ROCm 7.13.0 Preview (May 15, 2026) that extended support for MI350 and MI300 GPUs. These updates underscore ROCm’s role as a critical enabler of AMD’s data center ambitions.

Competitors like NVIDIA dominate the AI GPU market, but AMD is carving out a niche with its open software stack approach. By offering the ROCm Certification, AMD aims to accelerate skill development among developers, making its GPUs more attractive for enterprise-scale AI deployments. Given the $879 billion market cap of AMD and the growing adoption of its Instinct GPUs, this certification could serve as a meaningful signal to employers seeking AI talent.

What’s Next?

The timing of the certification program is strategic. AMD’s MI300 series accelerators are expected to play a key role in second-half 2026 deployments in AI clusters. Developers who earn the Associate certification now will likely position themselves as early movers in this expanding ecosystem.

For those attending the Advancing AI 2026 conference, AMD is offering in-person ROCm certification courses. Participants can work directly with AMD experts, complete lab exercises, and earn their certification badge on-site. This immersive experience is a compelling option for AI professionals looking to enhance their career prospects.

To get started, visit developer.amd.com.

Image source: Shutterstock Source