Problems in physics often inspire mathematical solutions, occasionally leading to the development of new mathematical objects. Mathematicians may then explore these constructs independently, sometimes uncovering new compelling physical interpretations in the process. This thesis contributes to this dynamic interplay between mathematical abstraction and physical reality, with a focus on algebraic curves. It aims to present findings that resonate with and are useful to both the mathematics and physics communitiesWe first explore the connections between algebraic curves and integrable systems, focusing on the KP equation, a nonlinear partial differential equation describing the motion of water waves. Our approach is based on the connection established by Krichever and Shiota, which showed that one can construct KP solutions starting from algebraic curves using their theta functions. This lead also to a new perspective on the classical Schottky Problem which has interested algebraic geometers for several decades. In this thesis, we explore KP solutions arising from curves which are not smooth, having at worst nodal singularities. We introduce the Hirota variety, which parameterizes KP solutions arising from such curves. Examining the geometry of the Hirota variety provides a new approach to the Schottky problem, which we study for irreducible rational nodal curves. We conjecture and prove up to genus nine a solution to the Schottky problem for rational nodal curves.When applying algebraic geometry or combinatorics to areas of physics such as integrable systems or particle physics, positivity, in particular the positive Grassmannian, plays a major role. In the last decade it has garnered much attention from physicists through its connection with scattering amplitudes, which can be computed as volumes of amplituhedra. An amplituhedron is the image of the nonnegative Grassmannian $\mathrm{Gr}_{\geq 0}(k, n)$ under a totally positive linear map $\tilde{Z}: \mathrm{Gr}(k, n) \to \mathrm{Gr}(k, k+m)$. In this dissertation we study Grasstopes: generalizations of amplituhedra in which we allow arbitrary linear maps. As a result, we give a full description of $m=1$ Grasstopes, recovering some results about $m=1$ amplituhedra, and introduce some new directions of study. Though so far the study of Grasstopes has been motivated by pure mathematical interest, one hope is that physicists may come up with a use for them as well.We continue to draw inspiration from particle physicists in our study of the positive orthogonal Grassmannian. We initiate the study of the positive orthogonal Grassmannian geometrically, for not necessarily maximal dimensions, and with varying signature coming from the quadratic form. In particular we prove that, for arbitrary signature, the positive orthogonal Grassmannian for $\OGr_{\geq 0}(1, n)$ is a positive geometry, confirming physicists' intuition.Finally, we highlight the value of computation in algebraic geometry by revisiting classical problems. The centuries-old uniformization theorem states that an algebraic curve is equivalent to a compact Riemann surface. However, connecting a Riemann surface to an algebraic curve utilizes Riemann theta functions, which are infinite sums of exponentials, so this classical equivalence is transcendental, leaving a divide between analytic and algebraic approaches. In this thesis we make a step in bridging this divide. We present an algorithm which uses discrete Riemann surfaces to approximate the Riemann matrix of any square-tileable translation surface. We apply our algorithm to specific examples of Jenkins-Strebel representatives, a dense family of translation surfaces, leading to several conjectures about their underlying algebraic curves. We also study two-dimenstional linear spaces of symmetric matrices, addressing questions motivated by algebraic statistics and optimization. These spaces have many properties determined by their Segre symbols, which also provide a stratification of the ambient Grassmannian.