Drug compounds can have off-target effects on cardiac ion channels that can alter heart rhythm and sometimes lead to sudden cardiac death. Automated ion channel screening can now detect to what degree a novel compound blocks particular ion currents. Data from screening with mathematical models of cardiac action potentials can help us to understand what combined effect a drug will have on overall cellular electrophysiology [1]. A recent proposal [2] is to use this approach with a confirmatory experimental whole cell measurement, and a promising cell type to help us explore these effects is the induced human stem cell-derived cardiomyocyte (SC-CM) [3]. These cells are more readily available than mature human ventricular cells and so can allow a higher-throughput testing of compounds. We present a methodology for tailoring the ion channel densities in an existing mathematical SC-CM model to individual cells from action potential voltage recordings, and quantify associated uncertainties due to beat-to-beat variation. We provide a proof of principle in silico study using simulated datasets. We use a hierarchical Bayesian approach to construct parameter distributions for how we expect an ‘average’ cell to behave, allowing us to make predictions of drug action for an individual cell with an associated probability. For most ion channel densities, we find that we can successfully infer the ‘correct’ distributions for each dataset. We also infer a higher-level distribution which describes how these ion channel densities were selected. We then detail how this might be extended to look at the case of drug effects on SC-CMs and how they compare with what we expect from ion channel screening. Once the combined effects of a drug on these cells is understood in terms of its action on ion channels, we hope to extrapolate this information into predictions in the adult human situation using mathematical models again.