Varying coefficient models allow for a more precise description of the relationships between independent and dependent variables as coefficients are allowed to vary. However, model selection with varying coefficient models can be a challenge, especially if model complexity is further compounded by structure in the data, such as clusters, which must be considered. Recently revitalized stagewise selection techniques perform model selection in a flexible way that can allow for such sophisticated models. Though a strong connection has been established between stagewise techniques and the popular penalized regression approaches, stagewise approaches have also been shown to be competitive in their own right. We present the Kernel Stagewise Estimating Equations (KeSEE) technique that provides computationally efficient model selection for varying coefficient models in the presence of longitudinal data.