We present a machine‐learning interatomic potential for Mo‐Si alloys based on the Atomic Cluster Expansion (ACE) formalism, validate its performance and apply it for studying interface phenomena. Structural parameters, elastic constants and melting temperatures of the crystalline bcc Mo, diamond Si and stable Mo‐Si alloys (Mo3Si, Mo5Si3 and MoSi2) are calculated and compared to experimental values. Using the trained potential defect formation energies are calculated and the thermodynamic stability of various MoxSiy alloys is discussed with focus on Mo3Si. Finally, the intermixing between Mo and Si phases is studied by performing interface simulations of Mo|Si. The crystallization behavior of the Mo3Si phase provides additional evidence for the off‐stoichiometric composition of this intermetallic phase.This article is protected by copyright. All rights reserved.