With the increasing complexity of technology and large quantities of data in our digital age, learning and training has become a major cost of employers. Employee competence depends more and more on how quickly one can acquire new knowledge and solve problems to meet pressing deadlines. This paper presents a practical method to use REU (Research Experience for Undergraduates) projects and crowdsourcing to help students learn complex content as needed. The major question addressed in this work is to find ways to reduce the cost of learning for both the learners and their mentors. As a work-in-progress, we would like to share our preliminary design in using natural language processing (NLP) and data mining tools to retrieve and structure instructional materials semi-automatically. [For Full Proceedings, see: ED562107.]