The search for new materials with tailored properties has long been a challenge due to the high computational and experimental costs involved. Inverse design approaches offer a promising alternative by enabling the creation of property-to-structure models, as opposed to the traditional structure-to-property model development. These approaches can overcome the limitations of conventional, funnel-like materials screening methods and accelerate the computational discovery of next-generation materials. In this talk, we will explore the application of graph neural networks (such as ALIGNN) and recent advances in large language models (like AtomGPT and DiffGPT) for both forward and inverse materials design, with a focus on semiconductors and superconductors. We will also discuss the strengths and limitations of these methods. Finally, materials predicted by inverse design models will be validated using density functional theory prior to experimental synthesis and characterization. The above projects are part of the NIST-JARVIS infrastructure available at https://jarvis.nist.gov/.