Preparing a New Generation of AI-Ready Researchers

The ALCF’s “Intro to AI” student training series has hosted over 700 participants from across the nation to date, helping to cultivate a new generation of AI researchers.

ALCF postdoctoral researcher Shilpika (center) leads a session on prompt engineering during Argonne’s Intro to AI training series. Seated around the table are Argonne colleagues Paige Kinsley (foreground, left), George Thiruvathukal (back, center) and Venkat Vishwanath (foreground, right). Image: Argonne National Laboratory

Since its launch in 2021, the ALCF’s annual Intro to AI-driven Science on Supercomputers training series has been helping students develop the expertise to harness the power of AI and HPC for scientific research.

Aimed at undergraduate and graduate students from U.S. universities and community colleges, the program offers a combination of hands-on sessions and expert guidance to teach attendees to effectively use AI in their fields. Over the past four years, the series has welcomed over 700 total participants.

ALCF's Marieme Ngom leads a session on the basics of using neural networks for scientific research.

Comprehensive AI for Science Curriculum

The primary goal of the series is to introduce students to using AI and HPC for science, equipping them with the skills needed to apply these technologies in their own research. The training covers fundamental AI concepts, while also giving participants the opportunity to work directly with ALCF’s powerful systems, such as Polaris and the ALCF AI Testbed. By providing this foundational knowledge, the series helps participants bridge the gap between theoretical AI concepts and practical application using world-class computational resources.

With weekly sessions led by ALCF AI experts, the program provides detailed instruction on topics like large language models, neural networks, model training, and AI accelerators. In addition, attendees learn about the infrastructure necessary to run large-scale AI models efficiently, from computational strategies to optimizing workloads on advanced hardware systems like GPUs and novel AI accelerators.

Each session concludes with a science talk, delivered by different domain scientists, offering insights into how various Argonne teams are using AI to drive advances across fields such as biology, cosmology, and materials science. These talks not only provide context for how AI is applied but also highlight the potential of supercomputing resources in accelerating science and engineering.

A key feature of the series is its hands-on learning approach, where participants apply their new skills using ALCF computing resources. Each session includes exercises that reinforce key concepts, while also providing a platform for students to explore AI applications in science. At the end of the program, participants who complete all exercises receive a certificate of completion and a digital badge, marking their achievements.

Expanding Reach and Impact

The ALCF training series continues to expand its reach, with over 250 students registered in 2024. By providing free access to advanced tools and resources, the program plays a crucial role in developing the next generation of AI researchers and practitioners.

In addition to live sessions, all training materials and session recordings are available on YouTube and GitHub, making them accessible not only to students but also to the public. This ensures that even those unable to attend in real time can access the curriculum and learn at their own pace, further extending the program’s impact and reach.

Through its commitment to education and workforce development, the ALCF training series is shaping the future of AI-driven science. By preparing students to harness the power of AI and supercomputing, the program is equipping the next generation of researchers to drive scientific breakthroughs and innovations.