Exploring nuclear physics through the fundamental constituents of the strong force, quarks and gluons, remains a formidable challenge. While lattice quantum chromodynamics offers the most promising first-principles framework for this pursuit, its practical implementation is arduous, especially due to the uncontrollable growth of quark-combinatorics, the so-called Wick-contraction problem of nuclei. In a recent work, we addressed this challenge by combining a machine learning-based optimizer with a novel randomized algorithm inspired by computational number theory, demonstrating that light nuclei, up to even Carbon-12, may be studied within the first-principles framework of lattice QCD. We demonstrated the efficacy of our methods by computing two-point correlation functions for Deuteron, Helium-3, Helium-4, and Lithium-7, achieving at least an order of magnitude improvement over existing algorithms with efficient implementation on GPU-accelerators. Additionally, our investigation revealed a remarkable and universal pattern: certain spin-color combinations consistently dominate the signal across all nuclei studied, hinting at a potential connection to an as-yet-unidentified symmetry in nuclei. Building on this development, in collaboration with the BNL lattice QCD group, we have initiated a program aimed at exploring binding mechanisms, matrix elements, and other nuclear processes of light nuclei at physical pion masses. In this talk, I will elaborate on our ongoing investigations and present some preliminary results from this effort.