This study presents preliminary work on using a data-driven non-linear surrogate model based on deep learning to generate realistic wind data for urban environments efficiently. We attempt to create accurate wind data for an urban environment using high-fidelity CFD data from Large Eddy Simulations (LES) and Convolutional Auto-Encoders (CAE) for non-linear surrogate modeling. The non-linear surrogate model extracts underlying non-linear modes from the high-resolution data snapshots, and the LSTM network trains on these specific modes. Modal predictions for future time steps are then obtained using the trained LSTM network similar to time-series prediction, without computationally expensive CFD simulations. We can decode these modes back to the physical space to get the relevant wind field data predictions. Since no prior information about the underlying governing equations is utilized for the projections, the method is entirely non-intrusive. One can easily extend it for other applications with minimal changes.