The affine Hilbert function is a classical algebraic object that has been central, among other tools, to the development of the polynomial method in combinatorics. Owing to its concrete connections with Gröbner basis theory, as well as its applicability in several areas like computational complexity, combinatorial geometry, and coding theory, an important line of enquiry is to understand the affine Hilbert function of structured sets of points in the affine space. In this work, we determine the affine Hilbert function (over the reals) of arbitrary unions of layers of points in a uniform grid (a finite grid with the component sets having equispaced points), where each layer of points is determined by a fixed sum of components for all the points. This extends a result of Bernasconi and Egidi (Inf. Comput. 1999) from the Boolean cube setting to the uniform grid setting. Our proofs follow a similar outline as that of Bernasconi and Egidi. However, there are two bottlenecks that arise in the uniform grid setting. We resolve these by using (i) a classical fact that a symmetric Jordan basis exists for the function space on a uniform grid, and (ii) an extension to multisets of an algebraic interpretation by Friedl and Rónyai (Discrete Math. 2003) of the notion of order shattering. The affine Hilbert function is, in fact, a stronger notion than the finite-degree Zariski closure, which is yet another important tool in the polynomial method toolkit. We conclude by giving an alternative proof of a combinatorial characterization of a variant of finite-degree Zariski closures, for unions of layers in uniform grids, obtained in an earlier work of the author (arXiv, 2021).