The hot deformation behavior of Nb–V–Ti containing high‐strength steel is studied using compression test. Based on the data generated during hot compression tests, a deep neural network (DNN) flow stress model is developed for the experimental microalloyed steel. The predicted flow stress by the DNN model is compared with flow stress predicted by conventional strain‐compensated Arrhenius constitutive equation. Results using the DNN model are found to be superior to the constitutive equation with overall correlation coefficient (R) greater than 99.8%. The accuracy of the developed DNN flow stress model finds to be significantly higher even at higher strain rates. A DNN roll force model is also developed for the plate rolling of the experimental steel. The predicted flow stress from the DNN flow stress model for the given rolling condition is also used as an input to the roll force model. It is found that DNN model is able to predict the roll force accurately by >98.4% as it takes care of various nonlinear process variable which cannot be accounted mathematically. It also validates the accuracy of DNN flow stress model.