Indoor air pollution has aroused increasing concerns due to its significant adverse health impacts. Indoor PM_2.5 exposure assessments often rely on PM_2.5 concentration measured at a single height, which overlooks the vertical variation of PM_2.5 concentrations accompanied by various indoor activities. In this study, we characterize the vertical profile of PM_2.5 concentration by monitoring PM_2.5 concentration at eight different heights in the kitchen and the bedroom, respectively, using low-cost sensors with high temporal resolution. The localized enhancement of PM_2.5 concentration in elevated heights in the kitchen during cooking was observed on clean and polluted days, showing dominating contribution from cooking activities. The source contribution from cooking and outdoor penetration was semi-quantified using regression models. Stratified source contribution from cooking activities was evident in the kitchen during the cooking period. The contribution in elevated heights (above 170 cm) almost tripled the contrition in bottom layers (below 140 cm). In contrast, little vertical variation was observed during other times of the day in the kitchen or the bedroom. The exposure level calculated using the multi-height measurement in this study is consistently higher than the exposure level estimated from the single-height (at 110 cm) measurement. A more significant discrepancy existed for the cookers (17.8%) than the non-cookers (13.5%). By profiling the vertical gradient of PM_2.5 concentration, we show the necessity to conduct multi-height measurements or proper breathing-height measurements to obtain unbiased concentration information for source apportionment and exposure assessment. In particular, the multi-height measuring scheme will be crucial to inform household cooking emission regulations.