This is a collaborative project to build an education AI expert system for question aswering from pdf.
Josephat Oyondi - Team Lead @ItsOyondi, Jean Baptiste @Baptiste1991, Elie Niringiyimana : @niringiyimanaelie
The Chatbot can efficiently answer questions and provide the sources from where the answer was extracted from.
We used Pinecone to create our Vector Database and OpenAI for Text and Image Embedding. The data was extracted from pdf documents processed the leading computer vision model (YOLOV7). We locally deployed the chatbot using Flask API.
We developed this chatbot to keep a record of the user queries and use them for further improving the knowledge base and make the bot more robust and able to answer all questions in the selected subjects. In the future, we can also improve the model to process multiple subjects spanning the education sector.
I am so proud of this team and the efforts they put in making my work easier as a Team lead to deliver a functional product.