This paper presents ElegansAI, a neural network model that leverages the connectome topology of the Caenorhabditis elegans to design and generate advanced learning systems. The objective of this approach is to integrate the intricate circuitry of biological neuronal networks into artificial ones, with the aim of exploring the advantages of incorporating bio-plausible connectome topology in deep learning models. ElegansAI outperforms randomly wired tensor networks, simulated bio-plausible networks, and state-of-the-art models such as transformers and attention-enforced autoencoders. The models achieve a top-1 accuracy of $99.99%$ on Cifar10 and $99.84%$ on MNIST Unsup in supervised image classification tasks and unsupervised handwritten digit reconstruction, respectively. The proposed method offers a unique approach to designing and generating connectome-inspired learning systems that harness the functional distribution of biological neuron circuitry. It is shown how bio-plausible structures integrated into artificial neural networks efficiently tackle complex tasks by evaluating evolutionary optimized neuronal motifs.