Imitation learning enables scalable transfer of complex manipulation skills to robots, but its effectiveness depends on high-quality demonstrations and robust policy learning, especially in dynamic, contact-rich environments. This thesis investigates how combining imitation learning with teleoperation and classical planners can teach dexterous manipulation across diverse real-world settings. We develop a teleoperation system for collecting demonstrations for bimanual, anthropomorphic robot hands (BiDex) and show that behavior cloning enables visuomotor policies for dexterous tasks like tool use. Building on this, we introduce a teleoperation system enhanced with force feedback (FACTR). With a dedicated curriculum, we show that imitation learning in this setting benefits from access to both force and visual modalities, leading to contact-rich behaviors with out-of-distribution generalization. We further demonstrate that classical planners can serve as effective teachers, and that teacher-student fine-tuning enables reactive motion generation with collision avoidance (Deep Reactive Policy). Lastly, we show that object-aware representations, via dynamic Gaussian Splatting, allow imitation learning from RGB alone. These results highlight how effective interfaces, architectural priors, and perception strategies enhance the reach of imitation learning.