My research focuses on integrating biological principles into neuromorphic systems. I have worked on developing adaptive, 'always-on' systems that operate on continuous data streams, leveraging mixed-signal neuromorphic hardware. My work involves a co-design approach, where analog electronic circuits are developed alongside computational models, ensuring compatibility and efficiency.

- Using a chip-in-loop paradigm, I trained a network with cross-homestasis to achieve stable attractor dynamics with the desired target activity. The trained spiking-RNN preserve long-term memory, and show paradoxical effect, soft WTA dynamics.

- Following a chip-co-design approach, we developed a chip that features a local calcium-based dendritic learning mechanism, integrated into a neuromorphic 'canonical' cortical circuit motif.

- I investigate the role of dendritic inhibition via SST cells in mismatch negativity and predictive coding.