Modern personalized medicine or precision medicine (PM) promises to take medical epistemology and evidence-based medicine a step further by accounting for variation among individuals. It promises tailoring of treatments to the unique biological characteristics of each person’s specific health problem through analysis of data from genomic and other molecular “omics” technologies, a widening array of monitoring technologies, as well as organoids and related technologies. In an important sense, it promises both personalization and scientific reliability, turning what has previously often been thought of as the art of medicine into a science. But how can treatments directed at the uniqueness of individuals be known and shown to work in an evidence-based world dependent on population-based knowledge? A main method for documenting that treatments work in PM is stratification, which is a population-based strategy that compares the patient to smaller, more similar, subgroups. But PM also sometimes promises to move beyond stratification: To treat the individual as a truly unique case without relying on population-based clinical studies. This chapter asks if and how this might be done. It is divided in two main parts: The first considers strategies that go into such personalization before a treatment is first given through an historically informed survey of different forms of PM. The second considers strategies that may allow personalization beyond stratification after the treatment is first given. The chapter ends by arguing that the art of medicine resurfaces amid promises of scientific, high-tech PM.