Amazon Web Services (AWS) has a new leader to run its high-performance computing GTM operations. Thierry Pellegrino, who is well-known in the HPC community, has taken on the role of Global Head of Advanced Computing at AWS, announced in a LinkedIn post.

“This position encompasses High-Performance Computing (HPC), Domain-specific Machine Learning (ML), IoT / Edge Computing, and Quantum Computing,” Pellegrino said in the entry.

Pellegrino previously was President at SMART Global Holdings (SGH), which in 2018 acquired Penguin Computing. Last month, SGH renamed itself Penguin Solutions to capitalize on the emerging market for AI server builders.

Pellegrino led the team at Penguin that helped Meta build out its generative AI infrastructure, according to his LinkedIn profile.

Before SMART, Pellegrino worked in various senior server infrastructure roles at Dell.

Now, he will join a group of talented HPC leaders that includes Ian Colle, Deb Goldfarb, Srinivas Tadepali and Brendan Bouffler (popularly known as Bouff), who are active participants in the HPC space and Supercomputing shows.

AWS is building “the world’s largest AI supercomputer in the cloud,” called Project Ceiba in collaboration with Nvidia. Project Ceiba will be hosted exclusively on AWS and will power Nvidia’s research and development efforts in AI. It will be powered by 20,736 Nvidia Blackwell GPUs and 10,368 Grace CPUs and deliver 414 exaflops of AI performance.

“That’s around 375 times more powerful than the current world’s fastest supercomputer Frontier,” AWS said on its Project Ceiba site.

AWS is trying to bridge the gap between scientific computing in industry and academia. Service providers such as Rescale are helping in the process by orchestrating cloud and on-premises workloads. RIKEN has also put a virtual version of its Fugaku supercomputer on AWS.

Pellegrino’s appointment comes as major cloud providers try to attract more HPC customers and their workloads into the cloud.

Amazon is offering a wide range of proprietary and open-source AI models, including Meta’s Llama and Anthropic’s Claude, via its Bedrock infrastructure. AWS also offers quantum computing technologies via its Braket service.

AWS has also taken steps to attract the resistant supercomputing crowd, many of whom have been reluctant to move to the cloud due to unjustified concerns regarding security and bandwidth.

AWS recently announced the commercial availability of AWS Parallel Computing Service (PCS), which the company has said will finally bridge the gap between HPC deployments in the cloud and on-premises.

“We can now feel that we’ve increased the level of ease of use to ensure that customers can more easily migrate their HPC workloads onto AWS,” Ian Colle, general manager for advanced computing and solutions at AWS, told HPCwire in an interview last month.

PCS removes that friction and customers can manage AWS HPC clusters the same way they manage on-premises environments. The key addition was the SLURM (Simple Linux Utility for Resource Management) scheduler, which allows customers to orchestrate their workloads, storage, networking, and other components in an HPC cluster.

AWS also offers HPC instances that are built on homegrown ARM-based Graviton CPUs and x86-based chips from Intel and AMD. The company also offers GPUs from Nvidia.

Pellegrino in his LinkedIn entry invited customers to talk to him.