Bhuwan Dhingra's computer science students will gain grant-funded practical experience in large language models. (Photo by John West/Trinity Communications)
New Grant Funds Hands-On Experience with Large Language Models at Duke
Students in Duke’s Natural Language Processing course are getting hands-on experience with cutting-edge artificial intelligence tools thanks to a teaching grant from Thinking Machines Lab, an AI research company focused on making advanced AI systems more accessible to researchers and students, to Assistant Professor of Computer Science Bhuwan Dhingra.
The grant is part of the company’s Tinker Research and Teaching Grants program, which provides funding and computing credits to universities using its platform to train and customize large language models. Through the program, instructors can receive $250 in computing credits per student for courses that incorporate the platform, as well as additional funding to support research and course development.
In Duke’s CS 572 course, the grant is being used to give students practical experience in “post-training” large language models — the process of refining and improving AI systems after their initial training so they perform better at specific tasks. Using the Tinker platform, students will apply these techniques in their final projects, gaining experience with the same tools used in modern AI research and industry. By supporting students and instructors, the program is designed to lower the barrier to working with advanced AI systems and to help train the next generation of AI researchers and engineers.