DOCUMENT INTELLIGENCE / PROJECT 01
RAG Document Assistant.
Multilingual answers. Grounded in your documents.
C# / .NET 8 · Semantic Kernel · Azure OpenAI
View repository ↗THE PROJECT
Real context.
Practical engineering.
A document assistant that answers questions from a private collection of government PDFs, built to help a relative navigate conflicting information about a Canadian humanitarian program.
Adapted from the project’s GitHub README. Explore the repository for implementation details, setup instructions, and source code.
Read the full documentation ↗THE APPROACH
01
Parse PDFs with PdfPig and divide the text into chunks.
02
Store embeddings in Qdrant for semantic retrieval.
03
Use Semantic Kernel and Azure OpenAI to generate answers grounded in the retrieved context.
WORKFLOW AT A GLANCE
PDFs → Embeddings → Qdrant → Grounded answers