CMT Sensei was a Final Year school project to develop an interactive AI chatbot built entirely on Open Source resources and run fully within the school's internal servers. The chatbot is trained on course material and general information about the school, allowing it to be used by both students in need, and school visitors.
Students can access the chatbot via a website via their student accounts, where the chatbot will selectively choose its reference material based on the current module the student needs help with. Rather than answer the student's queries immediately, the chatbot is designed to prompt the student with another question to further their thinking process and better their understanding, making it a useful learning tool.
Improved deployment steps on Docker to allow for 1-click deployment on any device.
Handled load balancing between the various chatbot systems (art generation, image scanning, etc)
Implemented the base framework for prompting the chatbot via restful api command on Unity and receiving an output back.
Created the Draw and guess minigame using the AI's image scanning feature.
Implemented a safety filter for the phototaking mode using the AI's image scanning feature to prevent inappropriate photos from being displayed.
Integrated all assets by the artists.
Frequently checked on the team to make sure they were on time, adjusting the schedule accordingly if they were having troubles with thir tasks,
Consistently updated FYP supervisor on progress and proactively worked in accordance to feedback given,
Created in-depth documentation for potential handover of the project in the future.