Researcher & PM (LLM) for a joint project with a research institute, Oct. 2024 ~

In a joint project with a research institute, a PoC on the application of LLM to healthcare-related big data was performed.

A major theme in this field is future prediction (of health conditions), but there are various hurdles to actual application. One of them is data confidentiality, which is where Local LLMs and Domain-Specific LLMs come into their own.

For a prototype of such LLMs, we developed a simple model that integrates multiple estimation methods to perform causal inference between items.

I was employed in the roles of both researcher and PM (Project Manager).

As a researcher, the formulation of the problem was somewhat complex due to the nature of the project; it was necessary to deal with various issues such as which of the vast amount of data to use, what criteria to use, how to evaluate, and what the Ground-Truth should be, in an appropriate manner for each inference method.

As a PM, I assigned tasks to several AI engineers and managed their progress.

In the end, we were also able to obtain insights for research issues (both medical and mathematical), which will be used as a basis for the next joint research project.