The present paper deals with regulation and cell divisionin a high-dimensional, nonlinear process governed by intricate biochemical signaling and physical constraints. Traditional mathematical models, primarily based on Ordinary Differential Equations often struggle to account for the inherent biological "noise" and the massive morphological datasets generated by modern microscopy. This study proposes a novel framework that integrates physics informed Neural Network with stochastic modeling to simulate the mitotic progression of human somatic cells in the present paper we utilize a hybrid approach combining Convolutional Neural Network for real-time automated segmentation of mitotic phases and Recurrent Neutral Network to predict the temporal evolution of cyclin-dependent kinase concentrations. The mathematical foundation rests on a modified reaction-diffusion system.