ShefGuide is an IBM project to help new students adapt to new environment faster by centralising all informations and datas into a Chatbot. This is a project that heavily emphasizes on team work, communication skills, problem-solving, and management skills, as staffs from IBM will monitor project progress and frequent reports are expected. The final chatbot comprises three parts: the large language model itself, the database(Scraped from university website), and the connector to GPU which connects the chatbot to the backend server and then to the users.

Kwun's team is the only team that sucessfully produced something, not only that, the final production is way over the scope of IBM's requirement of providing a demo, Kwun's team have an actual deployable application that can truly help new students adapt to new environment faster by centralising all informations and datas. Inspired by an academic paper “BloombergGPT: The First Corporate Specificed Model” which dives deep into how the first corporate specificed AI was made, by having two reinforcement learning models, one for common sense knowledges and the other for corporate specific knowledges.