The main purpose of this study was to explore how low-skilled worker’s learning activity influences skill improvement. Using a unique 2007 Human Capital Corporate Panel data-set from the South Korean manufacturing industry, we operationalize skill improvement over time among low-skilled workers. A worker is classified as ‘low skilled’ if he or she has a low education level and poor technical skills. Regression models show that low-skilled workers’ informal learning positively influences their skill improvement. In contrast, we note that supervisors negatively influence skill improvement of low-skilled workers when measuring the change in technical skill proficiency. Quality circle programmes also have a positive influence on skill improvement. In conclusion, skills can be improved through planned interventions that increase collaboration on the job. The results from this study help to highlight the importance of designing learning interventions for low-skilled workers that take account of their underlying education and skills.