Meriplex RAG-Based Documentation Chatbot
2023An AI chatbot using semantic search to retrieve information from internal company documentation.
Technologies Used
About This Project
This intelligent RAG (Retrieval-Augmented Generation) chatbot was developed to help employees find information across thousands of pages of internal documentation. The system uses a hybrid search approach that combines semantic vector similarity with traditional keyword matching to deliver results in under a second.
Built with .NET and Blazor, the application utilizes SignalR to stream AI generated responses in real time. It features an automated ingestion pipeline that processes new documentation and updates a ChromaDB vector store without manual intervention. The system also includes rate limiting and secure access controls to ensure responsible usage of the OpenAI API.
Key Achievements
- ▹Engineered a hybrid semantic and literal search strategy for high retrieval accuracy
- ▹Achieved sub second query response times with real time SignalR streaming
- ▹Automated the document ingestion and embedding pipeline for constant data updates
- ▹Implemented rate limited access controls to manage API costs and usage
- ▹Utilized Docker containerization for consistent deployment and scalability
Screenshots
Project screenshots coming soon