This semester, I decided to offer a series of seminars to faculty on leveraging large language models (LLMs) for technical tasks. Initially, I hoped to attract around a dozen participants. The response, however, far exceeded my expectations—135 faculty members signed up! The level of interest underscores the pressing need for educators and researchers to quickly master this new technology.
The seminars focused on applying LLMs to various professional activities, offering hands-on experience in the following areas:
- Seminar 0: Step-by-Step Creation of an AI Agent
This introductory session was designed for those curious about how to build AI agents using the APIs of GPT and Claude. Even without programming experience, participants were able to follow the steps and build a basic AI framework. The goal was to demystify the process and show that anyone can use AI tools with the right guidance. - Seminar 1: AI Agents in Instructional Activities
In this seminar, I demonstrated how LLMs can streamline various instructional tasks. Faculty learned how to use AI to generate lecture materials, grade assignments, and even create AI-powered tutoring systems tailored to specific lessons. This is a game-changer in education, as it allows teachers to focus more on interaction and less on routine tasks. - Seminar 2: Using AI for Grant Review and Preparation
This session showed how AI can assist with reviewing and drafting grant proposals. Participants saw examples of AI-generated proposals alongside actual panel reviews, learning how LLMs can align submissions with funding agency guidelines. This significantly accelerates the grant-writing process, especially for faculty juggling multiple research projects. - Seminar 3: Writing Manuscripts with AI
LLMs are not just for casual writing—they can assist in creating academic manuscripts. This seminar focused on how to use AI to draft complex academic papers, from structuring definitions and theorems to computational proofs. Faculty left with practical insights on using AI for technical writing while understanding its limitations. - Seminar 4: AI in Data Analysis
The final seminar dove into the technical application of LLMs for generating code and creating data analysis pipelines. A case study on analyzing complex transcriptomics data demonstrated the power of AI in automating intricate coding tasks, making it possible to quickly and efficiently develop reproducible analysis workflows.
LLMs like GPT, Groq, and Claude are becoming indispensable tools in both academia and industry. They can perform tasks ranging from coding to content generation with incredible precision. I have developed theories and technologies to enhance the use of these models in technical settings, ensuring a high degree of accuracy for specialized tasks. The thirst for this knowledge is real, as shown by the overwhelming response to these seminars. LLMs are not just tools; they represent a leap forward in productivity, innovation, and problem-solving.
