Everyone mentioning HIV attacking the immune system is incorrect. The problem is that HIV is incredibly mutable. We create many antibodies that bind to HIV, however HIV can mutate to escape them. There is a conserved portion of HIV (the part where it binds to CD4 T-cells) that a few antibodies called broadly neutralizing antibodies (bnAbs) are able to target. However, these bnAbs are relatively rare, and tend to happen in patients who have had HIV for a while.

It's incredibly difficult to generate these bnAbs from a vaccine. Your body generates tailored antibodies through a process of mutation and selection called affinity maturation. These HIV bnAbs require a ton of mutation. There is research on trying to shepherd B cells through many rounds of mutation and selection to create these bnAbs by vaccinating with multiple different antigens, but this is still a very difficult problem.

Another difficult vaccine is malaria. We have vaccines that create antibodies that help protect against malaria. However, having antibodies to a foreign protein diminishes our ability to further develop an immune response to said foreign protein. As a result, current vaccine candidates have struggled with creating a large enough antibody response to generate strong protection.

Source: I'm a soon-to-be Ph.D. graduate in a computational vaccine design lab.