In the western United States, there has been a significant increase in both the size and number of wildfires associated with the increasing drought conditions. The six largest wildfires in U.S. history have occurred in the last seven years. Four of the six fires were in California (2017 Tubbs Fire, 2018 Camp Fire, 2020 Bay Area Fire, and 2021 Dixie Fire), where the mountains create complex atmospheric flows that lead to terrain-induced wildfires from dry, downslope winds. To investigate the health outcomes associated with smoke exposure the spatiotemporal distribution of smoke plume concentrations is required. Air quality monitors are sparse and do not quantify pollutant concentrations solely from smoke plumes. To estimate smoke impacts, the monitoring data can be supplemented with information from satellites, wildfire emissions inventories, dispersion models, and chemical transport models. Results of air quality modeling efforts to simulate wildfire smoke plume transport in the western U.S. and estimate smoke exposure are presented here. The focus of this work is on two recent modeling efforts (1) a novel smoke exposure modeling framework that provides estimates by fuel type, fire size, and plume age and (2) gap-filling approaches that leverage machine learning algorithms to increase the usability of satellite aerosol remote sensing products for estimating wildfire smoke exposure.