Welcome

I’m a postdoctoral researcher in Neuromorphic Computing and AI at the Neuromorphic Software Ecosystems, Jülich Research Center, Germany. I aim to create efficient AI systems by borrowing from the brain’s computational principles. I approach this by integrating insights from biology, developing and testing models through computer simulations, and implementing these models in neuromorphic hardware. I am currently working on optimizing recurrent neural networks by integrating dendritic computing principles to improve activation sparsity and long-context understanding.

During my PhD at the Computational and Systems Neuroscience, Jülich Research Center, I developed a sequence learning model that offers a mechanistic explanation of how sequences can be learned and replayed in biological neural networks. The model incorporates nonlinear dendritic integration, Hebbian structural plasticity, and homeostatic regulation mechanisms that reflect key features of cortical computation. It provides an energy efficient sequence processing mechanism with high storage capacity by virtue of its sparse activity. In collaboration with the Electronic Materials Institute, Jülich Research Center, we implemented the sequence learning model on a neuromorphic hardware centered around memristive devices. The system could successfully learn and predict sequences with a low energy budget.