
Dr. Feiyi Wang
Dr. Feiyi Wang received his Ph.D in Computer Engineering in 2000 from North Carolina State University (NCSU). He worked as a principal research scientist at Microelectronic Center of North Carolina (MCNC), where he served as the lead PI and Co-PI for several projects by DARPA, the Defense Advanced Research Projects Agency.
In 2007, he joined Oak Ridge National Laboratory, and in 2019 was appointed Group Leader of the Analytics and AI Methods at Scale group. He has received numerous awards and is an ORNL Distinguished R&D Scientist. He currently serves as Group Leader for the Analytics and AI Methods at Scale Group.
Dr. Wang holds a Joint Faculty Professor position at University of Tennessee, and is a senior member of IEEE.
Congratulations to Feiyi Wang on being selected as an HPC-AI Vanguard.
Dr. Feiyi Wang
Distinguished Research Scientist
Group Leader
Oak Ridge National Laboratory
What was your first involvement in HPC or AI?
I joined the National Center for Computational Sciences (NCCS) and became involved with the Oak Ridge Leadership Computing Facility (OLCF) during the era of ”Jaguar,” a pioneering petascale supercomputer.
My first experience with the full lifecycle of a high-performance computing system came with “Titan,” when I participated in every phase — from the initial RFP and early technology evaluation to building statements of work, system acceptance, deployment, operation, and ultimately, decommissioning.
In 2019, I was appointed as the Group Leader of the Analytics and AI Methods at Scale group. Since then, we have fully embraced and pushed forward on the mission to develop, deploy, and leverage cutting-edge AI to drive groundbreaking scientific innovations on large-scale, world-class computing systems.
What is your passion related to your career path?
I am a computer scientist by training, with a deep-rooted passion for system R&D and data-driven operational intelligence at massive scales. High-performance computing and AI are both rapidly evolving fields, each with its own dynamic community and distinct needs. However, it is at the intersection of these two fields where I find the most exciting opportunities. The potential of combining HPC and AI to push scientific boundaries drives me to explore new ways of leveraging these technologies together to unlock breakthroughs that neither could achieve alone. It is no exaggeration to say that the fusion of HPC and AI fuels my career path, allowing me to work at the forefront of scientific innovation.
Do you prefer working as an individual contributor or a team leader?
I definitely prefer working as a team leader. HPC and AI, particularly AI for scientific applications, are highly interdisciplinary fields, and for any meaningful project, it’s nearly impossible for one person to handle everything alone. My current team reflects this diversity, comprising experts in computer science, computational fluid, computational physics, and cognitive science etc. Working together, we not only spark new ideas but also bring them to life in ways that none of us could achieve individually. Leading a team through the journey of taking an idea from its early stages to a fully realized outcome is an incredibly rewarding experience.
Who or what has influenced you the most to help you advance your career path in this advanced computing community?
Throughout my tenure at the DOE national lab, I have been fortunate to learn from and be guided by many colleagues and peers. In particular, I would like to express my gratitude to Dr. Gina Tourassi, former NCCS division director, and Dr. Arjun Shankar, my former section head and current NCCS division director, for their unwavering support. They consistently encourage me to step out of my comfort zone and take risks, knowing I have their full backing. I am also deeply thankful to Prof. Jack Dongarra, former director of UT’s Innovative Computing Laboratory (ICL), for his generous support of our internal HPL on the Frontier project and his heartfelt encouragement for my personal career growth.
What are your thoughts on how the nation can build a stronger and deeper pipeline of talented and passionate HPC and AI professionals?
HPC talent is highly sought after, and AI professionals are in even shorter supply. Those who possess expertise in both fields are truly rare unicorns, so to speak. While these aren’t groundbreaking ideas, we must address the shortage by starting with early exposure to these fields, engaging students through hands-on projects, mentorship and support networks. Universities should offer interdisciplinary programs that blend computational science with domain-specific applications, and foster partnerships between academia, industry, and national labs to provide real-world experience. By integrating these efforts, we can cultivate a skilled, passionate workforce ready to push the boundaries of HPC and AI.
What does it take to be an effective leader in HPC and AI?
Please take this as my aspirational goal, rather than the perfect answer for this: To be an effective leader in HPC and AI requires a balance of forward thinking and adaptability. AI is a fast-moving field, and with HPC and AI converging at a critical crossroads, staying ahead means constant learning and keeping up with new developments. Leading by example is crucial – demonstrating a commitment to innovation, while focusing on AI’s potential for advancing scientific discovery. It’s about fostering collaboration and driving progress in both fields, ensuring the team is always pushing boundaries and prepared for the next wave of technological breakthroughs.
What is the biggest challenge you face in your current role?
People, people, people – as I have alluded in my previous answer. HPC talent is such a valuable and rare resource. We are constantly striving to attract top-tier professionals, but we often lag behind the industry in both talent acquisition and securing necessary resources. This gap puts pressure on us to innovate with fewer resources while ensuring we stay competitive in the fast-evolving fields of HPC and AI.
What changes do you see for the HPC / AI community in the next five-10 years, and how do you see your own skills evolving during this time frame?
I foresee a stronger convergence between HPC and AI, with AI increasingly integrating into classical modeling and simulation, such as through surrogate models and AI assistants (or science co-pilot), that are capable of synthesis or even hypothesis generation and planning, built on domain-specialized foundation models. Mixed-precision computing may become a mainstay and an important consideration when developing new applications. However, challenges will intensify as the industry continues to outpace us in both talent and infrastructure.
The foundational models we are building on Frontier today will quickly become outdated as AI evolves at an unprecedented pace. We are still searching for the game-changing breakthrough that will make “AI at scale” truly impactful for science, and this quest will drive much of the field’s progress.
Additionally, quantum computing has the potential to disrupt the landscape, pushing us to adapt and prepare for the convergence of HPC, AI, and quantum technologies. My own skills will need to evolve accordingly – remaining agile, pushing the boundaries of scalable AI for science, and embracing new paradigms to stay competitive in this fast-evolving ecosystem.
Do you believe science drives technology or technology drives science?
I’m not much for philosophical pondering, but in my view, science and technology feed into each other—advances in one often spark breakthroughs in the other, creating a cycle of mutual growth.
Would you like to share anything about your personal life?
I developed a passion for film photography during my time at NCSU many years ago, and bought my first SLR camera, a Canon EOS Elan II. Since then, the rise of digital cameras and smartphones disrupted the photography landscape, turning me and my film camera into relics. I adapted by diving into flash photography and even set up a home studio, which is something a smart phone camera can’t easily replicate, and it is a lot of fun playing with light. But when my daughter and son arrived, they quickly saw the house as their playground. I adapted again—now all my camera and flash gear sit in the attic, gathering dust!