Development of targeted nanoparticle for personalized cancer therapeutics often requires complex synthetic schemes involving both supramolecular self-assembly and multiple chemical modifications. When attempting to encapsulate a novel drug structure, these processes can be difficult to predict, execute, and control. In my talk I will describe a new method to accurately and quantitatively predict self-assembly of kinase inhibitors drug molecules as well as chemotherapeutics into nanoparticles based on their molecular structure and properties. These drug nanoparticle assemblies are stabilized with a new kind of excipient with ultra-high drug loadings of up to 90%. Using quantitative structure-nanoparticle assembly prediction (QSNAP) calculations and machine learning, a new algorithm was developed as highly predictive indicators of both nano-self assembly and nanoparticle size with unprecedented accuracy. From the entire drug space of more than 6000 compounds, 290 drugs were identified as nanoperticle formers. This approach can be further developed to all types of nanoparticles and enable cargo and vehicle matching which will significantly facilitate drug formulation processes.