Effective problem-solving necessitates managing cognitive load to enhance learning and performance in contemporary data-driven contexts. This study examines how prompt engineering skills influence intrinsic, extraneous, and germane cognitive load during complex data analysis tasks. Utilizing Cognitive Load Theory (CLT), we performed a qualitative study with 15 university students and explored their experiences through semi-structured interviews. Findings reveal that prompt engineering reduces intrinsic cognitive load by structuring problem-solving processes, minimizes extraneous cognitive load by filtering irrelevant information, and enhances germane cognitive load by fostering schema development and strategic thinking, reducing the overall cognitive load. These results emphasize the need for fostering AI literacy, which significantly contributes to learning and problem-solving. Future investigations will build on these results through quantitative studies to assess variations in cognitive load across differing levels of prompt engineering proficiency.