The Atlantic Forest is the second-largest rainforest in the Neotropical region, considered a hotspot for conservation due to its high rates of endemism. Because of its complex geological history, environmental heterogeneity, ecological gradients, and species diversity and turnover, the rocky mountains of the Atlantic Forest are excellent systems to study the evolutionary and ecological dynamics of the Atlantic Forest. Pitcairnia flammea Lindl. (Bromeliaceae) is endemic to the Atlantic Forest, and occurs at an elevation gradient from sea level to more than 2000 meters above sea level, presenting high morphological differences and low gene flow among populations. In this study, we aim to understand how genetic drift and natural selection affect the genetic diversity of P. flammea along the altitudinal and environmental gradients of the Atlantic Forest Mountains. We obtained SNP markers derived from RAD-seq of >160 individuals from eight populations of P. flammea distributed along the Atlantic Forest altitudinal gradient. The SNP calling was performed using the reference genome of the species. Using this data, we will 1) describe diversity, effective population sizes, genetic structure, and gene flow among populations; 2) identify genomic outlier regions possibly under positive selection; and 3) reconstruct demographic variations over time. Our preliminary results evidenced varying levels of diversity (percentage of polymorphic loci; nucleotide diversity (π); observed and expected heterozygosity). The populations from the highest altitudes showed the highest levels of genetic diversity and lower inbreeding coefficients. Population structure with pairwise FST varied from 0.081 to 0.35 and was also higher for high-elevation populations, ranging from 0.27 to 0.35. In agreement, the neighbor-joining analysis revealed that the eight populations are clustered into two major groups, indicating closer connectivity between lowland and intermediate altitude populations compared to the high-elevation populations. The maximum likelihood analysis in ADMIXTURE revealed eight genetic clusters (K=8), each population as a single unit, except by two geographically close populations clustered together, and the highest-elevated population divided into two clusters. Together these preliminary results indicated high genetic structure and weak gene flow among populations, which suggests a strong genetic drift over P. flammea populations along the Atlantic Forest altitudinal gradient. Future analyses of outlier loci will allow us to evaluate the relative effect of natural selection and genetic drift on molding these populations' diversity. Additionally, demographic reconstruction will help to uncover the evolutionary history of P. flammea populations along the altitudinal gradient. Our study will help to understand the processes responsible for the high diversity and endemism patterns observed in the Atlantic Forest Mountain populations, and how they interact in molding this remarkable complex tropical forest.