Lack of skills is arguably one of the most important determinants for high levels of unemployment and poverty. Targeting youth unemployment and also important because of its strong influence on other important social outcomes. Using a “fuzzy” regression discontinuity design, we examine the employment effects of a vocational training program in Nepal launched in 2009 over a three-year period. We find program participation generated an increase in non-farm employment of 28 percentage points for an overall gain of 95 percent, three years into the program. The program also generated an average monthly earnings gain of 2,167NRs (≈ 29 USD) or 171 percent. Applying heterogeneous local average treatment effect (HLATE) estimators, we find striking differences in the impacts by gender: program impacts are almost double the size for women than for men.