CHARM GPT (Locally hosted LLM Model)

We have created a locally hosted AI using free open-source LLM models. Unlike ChatGPT, these are locally hosted in the Ferris network. The primary purpose of creating this is to ensure that none of the CHARM healthcare-based data is uploaded to the internet, and also to seek LLM-based support at work. This system has a web UI with multiple co-workers from the CHARM team who can connect to it, who are in the Ferris local network. This demo will walk you through multiple open-source LLMs (from Meta, Google, DeepSeek, etc.) running side by side and how we can utilize them for our own purposes. The long-term purpose is to create AI agents that can automate some of the manual interventions we perform at CHARM, i.e., cleaning messy healthcare data to obtain structured and standardized attributes.