Driving the innovations needed to bring fusion power to the grid
Engineering technologies that turn fusion concepts into real-world devices
Exploring the fundamental physics of the fourth state of matter
Understanding how fusion plasmas interact with, stress, and alter materials
Studying how matter reacts to extreme temperature and pressure
Turning breakthrough fusion and plasma research into practical technologies
News / Fusion energy / AI ignites a new Data Science Division at the PSFC
The Plasma Science and Fusion Center launched a new Data Science Division, led by Principal Research Scientist Cristina Rea, to advance fusion research through AI, machine learning, and high-performance computing collaboration with the PSFC and beyond.
Feeling social? Share this
There are a handful of fields that have yet to be transformed by the touch of AI, but fusion science is not one of them. Fusion researchers have been using AI and especially its subset, machine learning (ML), to analyze data, learn from it, and identify the most efficient path forward. In a 2025 Journal of Fusion Energy guest editorial, MIT Plasma Science and Fusion Center Principal Research Scientist Cristina Rea outlined how ML is enabling fusion simulations to run in a fraction of the time required by traditional models while still maintaining highly accurate predictions. AI and ML’s impact on fusion—especially the race to commercial fusion power—is already substantial, and it will only continue to grow.
To strengthen the role of advanced computing and AI across the PSFC’s fusion and plasma research programs, a new Data Science Division has been launched, with Cristina Rea appointed as the inaugural Division Head. Rea’s division will unify the Center’s data science efforts while creating new opportunities for collaboration, education, and innovation, within and beyond MIT.
According to PSFC Director Prof. Nuno Loureiro, “Data science, artificial intelligence, and machine learning are quickly becoming the most important computational tools of our time. At the PSFC we are fortunate to have Cristina, who is unanimously recognized by the fusion community as a thought leader on these topics. She was the obvious choice to lead this effort.”
Since its founding nearly 50 years ago, the PSFC has focused on multidisciplinary approaches, and the Data Science Division will connect expertise across the Center’s five existing research divisions, bringing new computational methods to bear on challenges ranging from plasma instabilities to materials development.
“Fusion research has entered a data-rich era where AI, machine learning, and digital engineering are essential to accelerate discovery and commercialization,” says Rea. “This new division positions the PSFC to transform massive experimental and simulation data into predictive insight for the experiment-to-pilot-plant transition.”
The PSFC is already a leader in the arena of AI and machine learning, in part driven by Rea’s Disruption Group and a portfolio that includes the Machine Learning Working Group, which Rea launched in 2018 to connect experts across institutions; the Open and FAIR Fusion initiative, which develops open-source tools and datasets for machine learning applications; and the Computational Physics School for Fusion Research (CPS-FR), which trains students and early-career scientists in high-performance computing and data science.
The Data Science Division will serve as a hub for training and collaboration in computational science, extending Rea’s work with CPS-FR and the PSFC’s longstanding role as an educational center for fusion researchers worldwide. Rea embodies the PSFC’s educational mission; her group currently houses eight postdoctoral fellows, six graduate students, and two undergraduates. “These are transformative technologies that are replacing our day-to-day paradigm of conducting research, augmenting our capabilities and accelerating progress towards our end goals,” notes Rea. “It’s essential that our students gain the hands-on skills they need to tackle these modern fusion energy challenges.”
Rea joined MIT in 2016 as a postdoctoral associate. Her research focuses on interpretable and adaptive AI methods for predicting plasma behavior, including real-time stability assessments in major international devices such as DIII-D and the European TCV. She also serves as the PSFC liaison to the International Atomic Energy Agency Collaborating Centre on AI in fusion and plasma science.
“These technologies are changing the pace of discovery,” Rea says. “They allow us to see patterns, test ideas, and make connections that were previously out of reach. It’s an exciting time to be part of that transformation.”