KRAS is one of the most frequently mutated oncogenes in human cancer. Despite more than three decades of research, indirect approaches targeting KRAS-mutant cancers have largely failed to show clinical benefit, and direct approaches have historically been stymied by the apparently ‘undruggable’ nature of KRAS. In this case study, I’ll describe efforts at Amgen to identify cysteine-reactive molecules capable of selectively inhibiting a prevalent mutated form of KRAS, KRASG12C. These efforts leveraged iterative screening and structural biology studies, property-based optimization, and careful process engineering to deliver a highly potent and selective covalent inhibitor of KRASG12C: sotorasib (LUMAKRAS®/LUMYKRAS®). I'll also describe structural insights into the allosteric inhibition of KRASG12C and their implications for inhibitor efficacy and selectivity.