This paper analyzes the effects of cognitive ability on job satisfaction. Using PIAAC data we explore whether job satisfaction predictors observed in the literature change their behavior once we control for cognitive ability, proxied by an explicit test of individual numeracy level. We show that both, learning at work and job autonomy affect differently the job satisfaction for workers interacting with their level of cognitive ability measured. While learning increases job satisfaction for highly cognitive-able workers, task discretion (a measure of job autonomy) is particularly well valued by lower cognitive-able workers.