Our project is focused on highlighting the benefits and pitfalls associated with utilizing Large Language Models (LLMs) for process automation.

Specifically, we are attempting to emulate the role of a hiring manager with ChatGPT and then measuring its accuracy, consistency, efficiency, and precision to determine its efficacy in processing candidates by their resumes. The goal is to show that this AI technology, while highly efficient, may not be suitable for many of the applications in which it is being utilized due to its easily manipulatable nature, its tendency to hallucinate, and its inconsistent results.

Overall, we just want to point out that this tool is far from perfect while shining light on the areas where it performs particularly well and areas where it performs poorly in hopes that it might be implemented in a more prudent manner going forward. Furthermore, we are trying to explain how and why certain errors tend to occur when using LLMs, generally, by examining this specific use case. 

Personally, I find this project important because I have seen the downside of such rapid, irresponsible implementation in this regard. Namely, I have seen how difficult it is for my peers to attain interviews with employers due to AI algorithms dismissing their resumes for seemingly mysterious reasons. I want to help make this process simpler and fairer for both sides. 

Jude Roberts '25

Mentor: Dr. Sherif Abdelhamid, Assistant Professor, Computer and Information Sciences
Major: Computer Science
Hometown: Cloverdale, Virginia