Breath analysis is an emerging research field with tremendous potential for advance personalized, non-invasive health screening and diagnostics,while new sampling instrumentation tools andanalytical detection methods are developed. Notwithstanding of the quick development of commercially and researcher-built experimental samplers, no robust and repeatable VOCs' profile technologies have been clinically validated. Such is due to lack of an optimal standard procedure for selective breath sampling which ends in a wide range of contradictory reported results. Challenges of most breath samplers are also related to the substances' concentrations that are source (oral cavity, oesophageal and alveolar) dependent and their low values (in ppbv - pptv range). Here, we present a suitable and novel technology for selectively sampling exhaled air regarding the subject's: age, gender, metabolic production of CO2, smoking habits, nutrition and health conditions. The technology was aimed to perform real time flow measurements and collect a pre-determined portion of exhaled air by synchronizing a previously modelled respiratory cycle with the breathing cycle of the user. Through real-time synchronization of breathing cycles, the system can detect optimized sampling instants by machine learning-based algorithm. A first set of tests was conducted to evaluate the robustness and efficiency of the software's sampling algorithm with two cohorts of participants (n=15 and n=30) with different age groups (2-5 years old and 18-27 years old, respectively). The ability to selectively differentiate exhaled air was also tested through collection and posterior analysis of oesophageal and alveolar air samples obtained from an independent cohort of university students (n=31). Although it requires instrumentation improvements and optimization the breath sampling technology coupled with an efficient analyser device (GC-IMS), the results herein presented suggest a promise step forward in breath sampling adapted to users' age, genre and physiological condition.