This paper investigates how workers' skills and job application behavior contribute to the gender wage gap using data from a leading online labor platform. We utilize machine learning models to quantify the value of workers' skills and estimate their impact on wages. We find a substantial raw gender wage gap of over 30% that can, however, be fully accounted for by three factors: differences in workers' skills, differences in the projects they apply to, and differences in asking wages. Our findings indicate no employer discrimination based on gender. Instead, the gender wage gap emerges because men and women seem to use the platform in different ways. Women prioritize consistent income, while men pursue higher-paying, occasional gigs. These differences likely stem from different constraints and labor market opportunities outside the platform. According to our results, the flexibility of the online gig economy is unlikely to favor women.