Research in education increasingly uses mixed data types, such as a combination of ordinal item scores, continuous response times, and discrete variables based on process data. Generalized linear latent variable models (GLLVMs) fitted for such data can be used to infer relationships between multiple latent constructs and their development over time. These models are often high-dimensional and require efficient estimation methods. We present a fast estimation method for GLLVMs and illustrate its application to longitudinal data and joint modeling of process and performance data. Future extensions to support multilevel models and further types of nonignorable missing data are described.