Fundamental to rational drug and material design is a clear understanding of underlying molecular mechanisms. However, these processes are often challeng- ing to probe experimentally. Molecular dynamics simulations and machine learning (ML) offer high temporal and spatial resolution that can complement experimen- tal techniques, providing novel insights to accelerate discovery. This dissertation investigates molecular mechanisms related to kinase inhibition and chitosan/chitin self-assembly, with implications for rational drug and material design. Chapter 1 introduces and discusses the theory behind the methodologies ap- plied throughout the dissertation. Chapter 2 and 3 develop and apply molecular dynamics methods to identify covalently druggable sites in ERK pathway kinases and elucidate the mechanism of action of MEK1 inhibitors. Chapter 2 presents a novel computational protocol for assessing cysteine druggability, integrating continuous constant pH molecular dynamics simulations with pocket analysis. This protocol is prospectively applied to cysteines in MEK and RAF kinases. The analysis suggests that the GK+3 cysteine in RAF kinases and the back-loop cysteine in MEK kinases are druggable, offering promising opportunities for targeted covalent kinase inhibitor design. In Chapter 3 long-timescale MD was applied to investigate the mechanism of action of two MEK1 inhibitors (avutometinib and cobimetinib) in terms of their ability to disrupt RAF-MEK1 heterodimerization. The simulations reveal that cobimetinib disrupts the dimer through ligand-mediated destabilization of the MEK activation loop (A-loop), weakening the interface. In contrast, avutometinib enhances dimer stability via strong hydrogen bonding to the A-loop and a polar head group that forms unique inter-protomer contacts. Chapter 4–6 investigate the self-assembly mechanisms of the second most abundant biopolymer chitin and its derivative chitosan. Chapter 4 applies de novo simulations to elucidate the temperature-dependent mechanism and polymorphism of chitin self-assembly. The analysis reveals that hydrophobic interactions drive increased assembly at elevated temperatures. In chapter 5 a similar approach is used to investigate the influence of acetylation on chitosan self-assembly. The re- sults demonstrate that increasing acetylation reduces solubility, while the pattern subtly alters hydrogen bonding and chain registry. Chapter 6, non-equilibrium MD simulations are then used to study the electric field-induced contraction of chitosan hydrogels, providing mechanistic insights with relevance to the electro-fabrication of chitosan-based functional material. These simulations reveal that the electric field induces dewetting transitions between amphiphilic sheets which underlie the observed contraction behavior. Chapter 7 focuses on a preliminary study toward the development of machine learning models to classify the chemical reactivity of acrylamide-based warheads, using experimentally derived half-lives as training data. Chapter 8 summarizes the conclusions and lessons learned in this dissertation work and suggests future direction of studies.