Motivation: Workplace training is an indispensable part of workforce skill development, but it is often provided irregularly and selectively in private enterprises. This is a missed opportunity given its potential to contribute to employee wellbeing and human capital development, and to boost firm performance. Purpose: We examine workplace training provision and private returns to training in Myanmar, a country with a severely constrained formal educational system and underperforming private sector. Methods and approach: We use linked employer-worker panel data from micro, small and medium manufacturing enterprises in Myanmar. Our regression framework includes controls for production factors and other standard enterprise and owner characteristics, including employee and owner skills, and unobservable time-invariant employee and firm heterogeneity. Findings: First, we document a low prevalence of training in manufacturing enterprises. Second, we find a wage premium associated with training of about 8%. Third, we show that workplace training is offered selectively to workers and when this is the case, the wage gap between trained and untrained workers doubles. Fourth, we find that prior training (in other firms) or certified training do not contribute to higher wages, indicating that training is largely firm-specific. Finally, we find that the wage returns to training are particularly high for women and among least educated workers. Policy implications: Considering the wage return as a lower bound of productivity gains from training, our results suggest that government support of workplace training programs in private enterprises could help increase performance of the manufacturing sector in Myanmar. High returns to training among the least educated workers indicate that training could improve the overall income distribution in the economy.