Scaling Up Computing Education in a Time of AI

Artificial intelligence is not only a leading driver of the U.S. economy, it is spurring scientific discoveries in fields such as medicine, protein design, materials science, hurricane modeling, and more. 

Through a series of academic courses, the Texas Advanced Computing Center (TACC) at The University of Texas at Austin is preparing students to work with AI and other leading technologies in an evolving computational landscape.

Over the last two decades, TACC has hosted a diverse portfolio of National Science Foundation-funded supercomputers for academic research to transform how researchers compute, discover, and innovate including:

• Horizon – the NSF Leadership-Class Facility system launching in 2026
• Ranch – also part of the NSF LCCF, Ranch is the largest academic data storage system in the nation
• Frontera – the fastest academic supercomputer in the U.S.
• Vista – optimized for AI workloads
• Lonestar6 – serving UT System via the University of Texas Research Cyberinfrastructure Portal (UTRC)
• Stampede3 – a scientific workhorse supporting thousands of U.S. researchers
• Jetstream2 – a flexible cloud-based computing resource

“We're not only here to build big computers, but to make sure people use them well,” said Dan Stanzione, associate vice president for research at UT Austin and executive director of TACC. “One great way to do this is by teaching students on our campus, as part of their formal education, the methods of large-scale computational science."

Under the Hood of Supercomputing

Stanzione co-taught the Fall 2025 course CSE 380: Tools and Techniques of Computational Science, alongside Lars Koesterke, a researcher in TACC's HPC Performance and Architectures group. The course provides a foundation of the hardware principles, programming languages, and operating system environment for getting the best performance out of supercomputers.

"Everybody now uses Python versus compiled languages. What’s more, coding and report writing is done more with AI. Everything is AI now," Koesterke said.

CSE 380 is a required course in the Computational Science, Engineering, and Mathematics (CSEM) graduate program at the Oden Institute for Computational Engineering & Sciences at UT Austin. 

A core component of the CSEM curriculum is the study of algorithms and numerical methods directly applied to the students’ fields of science. 

“This is the one course that teaches how to translate algorithms into efficient, practical code using real tools and programming languages," Stanzione said. "We focus on efficiency in the presence of AI tools, and how to implement that on a machine effectively."

Launching Pathways in Computational Engineering

As part of the Computational Engineering undergraduate program in the Cockrell School of Engineering at UT Austin, most students begin with an introductory course such as COE 301, which lays the foundation for advanced study. The next step might be to branch into COE 322: Scientific Computation, a class recently co-taught by TACC's Victor Eijkhout and Susan Lindsey. 

Eijkhout brings extensive authorship experience to the classroom, having written three widely used textbooks: Introduction to High Performance Scientific Computing; Parallel Programming for Science and Engineering; and Introduction to Scientific Programming in C++ and Fortran.

 "COE 322 teaches the next level of C++ and various 'software carpentry' tools to make them more productive," Eijkhout said. "“C++ is an evolving language. TACC offers this class as a prerequisite to the Parallel Computing class (COE 379L), and because very few departments offer programming classes.”

The Changing Computational Landscape

Over the past decade, research computing education has grown substantially in scope and impact. When TACC first began offering courses in the mid-2000's, the curriculum was grounded in core computational fundamentals — from command-line proficiency and text editors to C and Fortran programming to OpenMP and MPI, and a rigorous understanding of system architecture and libraries.

“Today, this foundation is still essential, but the ecosystem around it has grown," said Charlie Dey, a research computing advocate and educator at TACC. "Students are now working in Python and C++, using tools like Jupyter Notebooks and VS Code, interacting with and building APIs, and thinking about front- and back-end systems."  

“Topics like system administration, security, cloud computing, data analysis, and machine learning are now part of the conversation," Dey added. "Our coursework has evolved to reflect that reality, meeting students where they are, while still grounding them in the fundamentals that make advanced research computing possible."

AI in the Classroom

Responding to the rapid rise of AI, Dey co-taught the Fall 2025 course COE 379L: Software Design for Responsible Intelligent Systems with TACC’s Anagha Jamthe and Joe Stubbs.

“This course covers the responsible design and implementation of intelligent systems,” said Stubbs, a research scientist in TACC’s Cloud & Interactive Computing group. 

Intelligent systems refer to computer-based systems that can perceive information, learn from data, make decisions, and act in ways that resemble human reasoning or problem-solving, often at scale and with some degree of autonomy.

 “COE 379L provides students with a deep dive into designing, implementing, validating, and operating real-world intelligent systems using scalable data analysis and modern machine-learning techniques,” he said. 

Each lecture is interactive combining presentations with discussion and hands-on exercises. Students apply what they have learned through guided projects and open-ended design challenges that mirror real engineering practice.

“The technical components are paired with a strong focus on the ethics and responsibility of AI development, including data bias, reproducibility, model maintenance, and evaluating models beyond simple accuracy scores."

Stubbs also teaches COE 332: Software Engineering and Design, a popular class he and his group created in 2018, which covers the methods and tools for planning, designing, implementing, validating, and maintaining large software systems. A required design course in the Computational Engineering program, it equips students with hands-on experience in distributed systems and cloud computing technologies.

"Throughout the semester, students complete a sequence of assignments that progressively integrate to form a complete, scalable distributed system for analyzing real-world datasets," Stubbs said.

From Machines to Meaning: How All the Pieces Come Together 

Future courses in 2026 and beyond will cover topics including artificial intelligence, engineering computation, parallel programming, software engineering, research computing in Life Sciences, and cybersecurity.

“The goal is for students to understand how advanced computing systems work from the machines we build, manage, and maintain to the tools, programming languages, libraries, and software that run on them," Dey concluded. "But just as importantly, I hope students see how all the pieces fit together. When students understand the full stack, advanced computing becomes more than technology, it becomes a way to enable scientific discovery, solve real-world problems, and change the world.

Learn more about TACC’s academic courses offered in 2026: Academic Courses

Article adapted from UT Austin's Texas Advanced Computing Center (TACC)