For many complex traits, gene regulation is likely to play a crucial mechanistic role. How the genetic architectures of complex traits vary between populations and subsequent effects on genetic prediction are not well understood, in part due to the historical paucity of GWAS in populations of non-European ancestry. We used data from the MESA (Multi-Ethnic Study of Atherosclerosis) cohort to characterize the genetic architecture of gene expression within and between diverse populations. Genotype and monocyte gene expression were available in individuals with African American (AFA, n = 233), Hispanic (HIS, n = 352), and European (CAU, n = 578) ancestry. We performed expression quantitative trait loci (eQTL) mapping in each population and show genetic correlation of gene expression depends on shared ancestry proportions. Using elastic net modeling with cross validation to optimize genotypic predictors of gene expression in each population, we show the genetic architecture of gene expression for most predictable genes is sparse. We found the best predicted gene in each population, TACSTD2 in AFA and CHURC1 in CAU and HIS, had similar prediction performance across populations with R2 > 0.8 in each population. However, we identified a subset of genes that are well-predicted in one population, but poorly predicted in another. We show these differences in predictive performance are due to allele frequency differences between populations. Using genotype weights trained in MESA to predict gene expression in independent populations showed that a training set with ancestry similar to the test set is better at predicting gene expression in test populations, demonstrating an urgent need for diverse population sampling in genomics. Our predictive models and performance statistics in diverse cohorts are made publicly available for use in transcriptome mapping methods at https://github.com/WheelerLab/DivPop.