Cryptococcal meningitis is the second most common cause of death in people living with HIV/AIDS, yet we have a limited understanding of how cryptococcal isolates change over the course of infection. Cryptococcal infections are environmentally acquired, and the genetic diversity of these infecting isolates can also be geographically linked. Here, we employ whole genome sequences for 372 clinical Cryptococcus isolates from 341 patients with HIV-associated cryptococcal meningitis obtained via a large clinical trial, across both Malawi and Cameroon, to enable population genetic comparisons of isolates between countries. We see that isolates from Cameroon are highly clonal, when compared to those from Malawi, with differential rates of disruptive variants in genes with roles in DNA binding and energy use. For a subset of patients (22) from Cameroon, we leverage longitudinal sampling, with samples taken at days 7 and 14 post-enrollment, to interrogate the genetic changes that arise over the course of infection, and the genetic diversity of isolates within patients. We see disruptive variants arising over the course of infection in several genes, including the phagocytosis-regulating transcription factor GAT204. In addition, in 13% of patients sampled longitudinally, we see evidence for mixed infections. This approach identifies geographically linked genetic variation, signatures of microevolution, and evidence for mixed infections across a clinical cohort of patients affected by cryptococcal meningitis in Central Africa.