A model-thoeretic version of Tarski's theorem

Working in any model theoretic structure, we single out a class of definable bipartite graphs that admit definable, close to perfect matchings. We use this result to prove a strengthening of Tarski’s theorem for the definable setting.