Coacervates can be formed via liquid-liquid phase separation when two oppositely charged polymers are mixed under specific conditions. These coacervates mimic the viscosity and density of natural cellular environments, creating a platform for early-stage drug screening. This project investigated the phase behavior and protein-partitioning capabilities of coacervates. To investigate phase separation behavior between carboxymethyl amylose and quaternary ammonium amylose, a phase diagram was created using an automatic microfluidic system. The system utilized machine learning to choose which sample concentrations to test. The results revealed a linear trend within a certain concentration range that indicates the coacervate charge ratio plays an important role in the occurrence of phase separation. Furthermore, confocal microscopy was used to demonstrate the successful partitioning of the protein into the coacervates. Future work will look at how to stabilize or inhibit protein-protein interactions using molecular glues within the coacervate.