Dec 2023
Kanojo AI was my first foray into the world of AI engineering, RAG (Retrieval-Augmented Generation), and prompt engineering. It was a project that not only piqued my curiosity but also expanded my understanding of what's possible with artificial intelligence.
The goal of Kanojo AI was to create a virtual assistant that could provide a more personalized and interactive experience. We started with basic features like creating new GPT assistants. Users could chat with these assistants, and each account was linked to its own unique assistant, ensuring a personalized experience.
As the project evolved, we began exploring more advanced features. We considered connecting to Stability Diffusion (SD) for generating profile pictures and even creating animated avatars. The idea of video and voice call functions was also on the table, along with the possibility of integrating OpenAI's text-to-speech API to give the assistants a voice.
Throughout this project, I learned a great deal about AI engineering and prompt engineering. Interacting with back-end data using database such as MongoDB and designing the sample layout for the user interface were new challenges that pushed me to think creatively and develop new skills.