The thiamine diphosphate‐dependent enzyme 2‐succinyl‐5‐enolpyruvyl‐6‐hydroxy‐3‐cyclohexene‐1‐carboxylate synthase (MenD) catalyzes a Stetter‐like 1,4‐addition of α‐ketoglutarate to isochorismate in the biosynthesis of menaquinone (vitamin K). Here, we describe the carboligation potential of MenD from Bacillus subtilis (BsMenD) for the nonphysiological 1,2‐addition of decarboxylated α‐ketoglutarate (succinylsemialdehyde) and various benzaldehyde derivatives. Furthermore, we engineer BsMenD variants for the enantiocomplementary asymmetric synthesis of functionalized α‐hydroxy ketones. Wild type BsMenD shows an excellent chemo‐ as well as high (R)‐selectivity for the carboligation of α‐ketoglutarate as the donor, and different benzaldehyde derivatives as acceptor yielding (R)‐α‐hydroxy ketones with up to >99 % ee. By engineering (S)‐selective BsMenD variants, based on the recently developed S‐pocket concept, we provide access to most of the corresponding (S)‐α‐hydroxy ketones with up to 98 % ee. In particular, benzaldehyde and meta‐substituted derivatives were converted with high enantioselectivities (ee of 91–98 % (S)). The significantly higher (S)‐selectivity of BsMenD variants than recently published MenD variants from Escherichia coli, could be attributed to a second‐shell residue next to the S‐pocket. A glycine residue, adjacent to the major S‐pocket residues I476 and F477 (standard numbering), is assumed to result in higher structural flexibility in the S‐pocket region of BsMenD, which in turn could result in improved stabilization of the antiparallel orientation of the acceptor.