The project aim to improve CFD prediction in built environment simulations using deep learning models. The main challenge of CFD simulation is the heavy computational effort required for higher fidelity simulations such as Large Eddy Simulations (LES) or Direct Numerical Simulations (DNS), which makes them practically inconvenient or even unfeasible in numerous applications. This project aim to solve this challenge through the development of a deep learning framework to speed up Large Eddy Simulation (LES) of airflow in built environments. The study consists of the training of a Convolutional Neural Network (CNN) using instantaneous velocity snapshots obtained from LES simulations from built environment domains. The framework's objective is to accelerate LES simulations by reaching statistically steady state faster, resulting in significant speedups in computational cost and simulation run-time while maintaining similar accuracy levels to standard LES simulations. This approach can greatly reduce the computational effort required for LES simulations in large scale environments. As main limitations to address is the challenge to extrapolate properly to different flow fields than the training set and understand the 3D turbulence structures.