As the only currently known family of ion channels that are directly gated by light, Channelrhodopsins (ChRs) allow for direct fluorescence imaging of voltage changes, an attractive method for studying neuronal circuits and heart muscle, and a useful complement to traditional electrode-based methods or calcium imaging. In this research project, we aim to engineer high performing rhodopsin proteins with the tools of machine learning. Using Archaerhodopsin-3 as our parent protein, our goal is to generate a library of new variants through SCHEMA algorithm and predict their functionality via classification and regression models.