Theoretical physics research poses a challenge for engaging undergraduate student participation. Undergraduates usually have not developed the mathematical and computational skill sets needed to fully appreciate the complexities of the research. Nevertheless, the students are often able to make valuable contributions through data collection, analysis, and modeling. In the process, they develop foundational skills for their future while helping to advance the larger goals of the theoretical effort. Several recent examples in the context of molecular astrophysics will be reported in this talk. A rigorous theory of molecule formation will be described and related to the process of star formation. Machine learning methods which aim to provide state-resolved collision rates for modeling non-LTE astrophysical environments will be presented, and contributions from undergraduate students will be highlighted.