Walk us through a time an experiment failed. How did you troubleshoot the hardware versus the data analysis?
