My research explores the computational and learning capabilities of biological neurons and networks, focusing on understanding input-output relationships in biologically realistic neurons and circuits. I introduced the Functional Circuit Index (FCI) to measure functional complexity, demonstrating greater complexity in human neurons compared to rodent neurons. Additionally, I developed a deep learning-based algorithm to optimize synaptic weights in detailed neuron models, revealing that single neurons can solve inherently complex classification tasks, such as high-dimensional parity problems. My broader interest lies in uncovering how single-neuron properties influence network-level computation and inspire brain-like AI architectures.