By using machine learning, the programming team was able to use data gathered from laboratory experiments to calibrate the sensor.
In two-week sprints, we developed the software package and user interface, received feedback from doctors, then iterated until the company was happy with the final product.