What is this?

This is a research study that I completed during my past user research internship at Red Hat (summer 2024).

The purpose of the study was to assess the current pain points that users have when interacting with an artificial intelligence model-serving interface.

 

What did I learn?

This testing required me to gain technical knowledge of how to serve LLMs, as well as how to run RAG applications on a computer. A central takeaway that I had from this testing is that the user's pre-testing technical knowledge heavily influenced how they interacted with this product. For example, those who frequently use Github did not struggle to understand what was being asked when the an input text field prompted for a Git DCO sign-off. Opposedly, those without technical knowledge found this to be one of their largest pain points. This finding validated the decision to conduct preliminary interviews before testing with our users to understand their technical experience with AI models and software development.

​

Please browse through this slide deck to view my research process, from when the problem was presented to the final recommendations given to stakeholders.