Fitness landscapes are a paradigm to understand how evolution proceeds. Yet such landscapes are only implicit about ecology and the diverse underlying communities in which organisms naturally thrive. In this talk, we show two concrete examples of how being part of an interacting community fundamentally alters evolutionary trajectories. Using bacterial communities as model systems, we show in one case how a species evolves in response to antibiotics depends on its interactions with another strain. In fact tuning this interaction can allow us to control how and whether antibiotic response evolves in the same environmental conditions. In the other case we show that sulfur cycling "pink berries" comprising diverse bacterial communities show remarkably slow down co-evolution due to strong mutualistic interactions in the community. In both cases, we show that simple mathematical models of community eco-evolutionary dynamics can capture our observations, and make new testable predictions. Our work highlights how ecological interactions can control evolutionary trajectories in bacterial communities.