Genetically encoded fluorescent biosensors are powerful tools for monitoring biochemical activities in live cells, but their multiplexing capacity is limited by the available spectral space. We overcome this problem by developing a set of barcoding proteins that can generate over 100 barcodes and are spectrally separable from commonly used biosensors. Mixtures of barcoded cells expressing different biosensors are simultaneously imaged and analyzed by deep learning models to achieve massively multiplexed tracking of signaling events. Importantly, different biosensors in cell mixtures show highly coordinated activities, thus facilitating the delineation of their temporal relationship. Simultaneous tracking of multiple biosensors in the receptor tyrosine kinase signaling network reveals distinct mechanisms of effector adaptation, cell autonomous and non-autonomous effects of KRAS mutations, as well as complex interactions in the network. Biosensor barcoding presents a scalable method to expand multiplexing capabilities for deciphering the complexity of signaling networks and their interactions between cells. Display omitted •Large numbers of fluorescent biosensors can be concurrently tracked in barcoded cells•Biosensor activities are synchronized in mixed populations of barcoded cells•Deep learning models facilitate image analysis for biosensor barcoding•Simultaneous biosensor tracking reveals signaling network structures and interactions Genetically encoded barcodes that uniquely identify cells expressing a particular fluorescent biosensor enable the simultaneous use of a broader set of biosensors in mixed cell populations and dynamic imaging of many cellular responses in a single experiment to reduce inter-experimental variability when interrogating complex biological interactions, such as the downstream consequences of receptor tyrosine kinase activation, with the help of image-based deep learning models for automated barcode unmixing.