The development of advanced polymer morphologies, involving nanofibers and polymer particles at the nano- and microscales, holds significant potential for emerging technologies in sensing, tissue scaffolding, filtration, and drug delivery. Among the techniques enabling the scalable synthesis of polymer films, chemical vapor deposition (CVD) stands out due to its solvent-free process and ability to provide conformal coatings on various substrates. When templated by liquid crystal (LC) films, CVD serves as a versatile approach for advanced manufacturing of nanostructured polymer morphologies. However, the full spectrum of morphologies and properties of the resulting polymer structures accessible through CVD-in-LC remains unrealized, owing to a limited understanding of key aspects of the process. This thesis presents a comprehensive investigation of novel polymer structure synthesis, ranging from nano- and microscale particles to ordered arrays and interconnected sheets of nanofibers, using two vapor phase polymerization techniques: chemical vapor polymerization (CVP) of paracyclophanes and initiated chemical vapor deposition (iCVD) within nematic LC templates. First, we report the surprising formation of a quasi-2D nanoporous membrane composed of interconnected, amine-functionalized nanofibers via CVP-in-LC. Unlike previously reported substrate-attached nanofibers oriented along the LC director, these networks form at the LC–gas interface via an interfacial polymerization-driven phase separation pathway. Insights obtained from molecular dynamics simulations reveal the role of precursor chemistry in defining the nanostructure of the polymer product. Second, we extend the LC-templated synthesis to iCVD using a cross-linkable monomer, resulting in the formation of polymeric nanospheres, microgels, and microspheroids. We elucidate a novel synthesis pathway and identify key energetic factors that govern morphological evolution. We further demonstrate the formation of singly dispersed, non-spherical microparticles that self-organize into hexagonal arrays within the LC film. Finally, we introduce a physics-informed deep learning workflow that enhances the resolution of optical microscopy images of polymer microparticles acquired in situ during iCVD-in-LC synthesis. This approach enables high-throughput microparticle size characterization without relying on costly, time-intensive, high-resolution ex situ imaging. Altogether, this work establishes a framework for the LC-templated synthesis and real-time characterization of functional polymer nanostructures, opening new avenues for programmable materials design at the nano- and microscale.