I created the Neuromorphic Physics team at the Laboratoire Albert Fert (a joint unit between CNRS, Thales and the University of Paris-Saclay), where my research focuses on advancing spintronic and neuromorphic technologies to build ultra-low-power Artificial Intelligence (AI) hardware.
Early in my career (2001–2009), I worked in spintronics, exploring the interactions between spin currents and magnetization. I was among the first to demonstrate the manipulation of magnetization in nanodevices using spin torques—an essential milestone in the development of Spin-Torque Magnetic Random Access Memories (ST-MRAM) that earned me the Jacques Herbrand Prize of the French Academy of Sciences.
Subsequently (2009–2016), I turned to neuromorphic computing, seeking to replicate the brain’s energy efficiency in artificial systems. In this context, I invented and demonstrated with my colleagues spintronic and ferroelectric memristors—ion migration-free artificial synapses that offer superior stability and power efficiency . The recognition of this work lead to my election as Fellow of the American Physical Society.
In (2017–2018) my collaborators and I demonstrated that nano-oscillators could perform real-time pattern recognition on spoken digits and vowels . This achievement showed how extremely small spintronic devices can execute complex neuronal computations by exploiting their inherent physical properties. It started a vibrant community in spintronic and physics-based neuromorphic computing—a field that continues to expand today. My contributions have been recognized with honors such as the CNRS Silver Medal, IEEE Fellowship, the IEEE Magnetic Society Mid-Career Award, and the Irène Joliot-Curie Prize of the French Academy of Science.
Today, my team and I are exploring how physics, machine learning and neuroscience can be brought together to build physical neural networks with enhanced computational capabilities and reduced energy consumption compared to current AI hardware – by up to a factor 1,000. The applications of our work are embedded AI for critical applications such as autonomous vehicles and medical diagnostics, providing low-cost, private on-device intelligence without relying on resource-hungry supercomputers. By combining cutting-edge AI algorithms with physics, we contribute to establish a transformative hardware platform for sustainable, high-performance computing—one that democratizes AI access worldwide, akin to the internet’s early development.
All my publications can be found here.
Beyond the lab, I like to engage the public through conferences, media appearances, a TEDx talk, and an illustrated book explaining AI fundamentals to children. I also enjoy creative writing, including poetry, as well as painting.
The illustrations below were realized by Sacha Berna with Julien Bobroff:
