Americas HPC Collaboration: Fostering an HPC Education & Workforce Development Ecosystem

High-performance computing (HPC) is now indispensable to disciplines ranging from climate modelling to drug discovery, yet the pipeline of professionals who can program, optimize, and operate HPC systems remains thin. The SC25 Americas HPC Collaboration Education & Workforce Birds-of-a-Feather (BoF) aims to launch an initiative to address that skills gap by convening organisations already running successful hands-on training programs across the continent to build a shared roadmap that enables scaling of these efforts.
The BoF will feature representatives from various international institutions, including the Advanced Computing System for Latin America and the Caribbean (SCALAC), the Latin American Advanced Networks Cooperation (RedCLARA), the International Centre for Theoretical Physics (ICTP), as well as national representatives from Colombia, Chile, Brazil, Mexico, Costa Rica, Canada, United States, and others.

• LATAM HPC Summer Schools – Brazil’s Santos Dumont School, Colombia’s CyberColombia series, and Mexico’s CONACyT initiative have jointly trained more than 900 participants in MPI, OpenMP, CUDA, and AI pipelines since 2021. Each school donates intake slots to under-represented institutions and streams every lecture over RedCLARA’s backbone for real-time remote labs.
• DevOps for HPC School @ CARLA – an annual week-long clinic that covers container orchestration, CI/CD, and performance engineering on heterogeneous clusters. Alumni now maintain production deployments at eight universities and two oil-and-gas research centres.
• Hackathons to solve real problems: focused events pairing US–LATAM teams with mentors from DOE national laboratories and SCALAC sites. Projects have delivered GPU-accelerated atmospheric down-scaling that runs faster on nodes donated by the Argonne Leadership Computing Facility and Chameleon Cloud.
• SCCAMP, The Supercomputing and Distributed Systems School, is a summer school and nonprofit event focused on Super Computing and Distributed Systems. It offers a series of courses centered on High Performance Computing, with a significant emphasis on practical sessions (more than half of the time). It targets Master’s and PhD students in Computer Science, Engineering, and other fields that could benefit from HPC.
• NVIDIA DLI It is a program developed by NVIDIA that provides hands-on training in AI, data science, and accelerated computing through courses, workshops, and certifications for developers, researchers, and educators. It offers practical experience with datasets and GPU tools such as CUDA, RAPIDS, and TAO. DLI supports workforce development, academic use, and worldwide access to high-performance computing and AI education.

What the BoF will deliver

Drafting a Chapter on HPC Education – Participants will an agreement that establishes the Americas HPC Education Chapter with two founding blocks:

• USA – represented by ACCESS and DOE national laboratories
• Latin America & the Caribbean – represented by SCALAC
• Europe and the World, – represented by ICTP- Trieste

This chapter should have four annual coordination meetings featuring country-level presentations on successful practices and engagement strategies. Focus on attracting new members and fostering regional collaboration. Each session will define key enablers, document replicable models, and align national priorities. The initiative will also develop a shared repository of training materials and allocate compute resources to support extended hands-on training events

Actionable roadmap – A twelve-month timeline will be proposed to:

• Gather bilingual (English/Spanish) course materials publicly available under a new area of the Americas HPC Collaboration website, explicitly created for the Education & Workshop chapter;
• Pilot a cross-listed graduate course, co-taught simultaneously from Oak Ridge, Argonne, Brazil, and Colombia, that awards joint credits.

Subsequently, the AHPCC Education & Workshop Chapter will be formally established with interested attendees, followed by an open discussion designed to outline action items that the community envisions to establish a robust flexible and inclusive workforce development pipeline . The chapter will meet every quarter to create a common archive of training materials and syllabi, and form a working group that focuses on identifying evolving HPC training resources, current needs of participants, and any gaps that the Chapter can help address.