A retrieval-augmented chatbot that responds using relevant source documents, developed as an applied knowledge-system project.
The problem
Teams need access to information buried in documents, but generic chat responses are difficult to audit.
The approach
Retrieve relevant source passages, ground responses in that context and assemble the application using Python services, a document index and a web interface.
Scope and limits
Portfolio research project. Project context does not imply the named organisation endorses or operates this site.
