Skip to content

Agentic-RAG — Retrieval-Augmented Generation with Semantic Kernel

** GitHub Repository:** Agentic-RAG

Project Overview

Tech Stack: Jupyter Notebook, Python, Semantic Kernel, ChromaDB

About: Agentic RAG demonstrates the use of Retrieval-Augmented Generation (RAG) with Semantic Kernel and ChromaDB. The project showcases how to build an AI agent that retrieves travel documents, augments user queries with semantic search, and streams detailed travel recommendations. It is designed to show real-world RAG applications, especially in the travel domain.

Technical Implementation

Core Technologies

  • Language: Python with Jupyter Notebooks for interactive development
  • AI Framework: Semantic Kernel for agent orchestration
  • Vector Database: ChromaDB for semantic search and retrieval
  • Domain Focus: Travel recommendation and document retrieval

Key Features

  • ChromaDB Integration: Factual retrieval from travel document collections
  • Semantic Plugins: Advanced search capabilities using semantic similarity
  • Travel Query Support: Specialized handling of travel-related inquiries
  • Streaming Chat History: Context-aware conversations with memory
  • Agent Orchestration: Coordinated workflows between retrieval and generation

Context & Impact

  • Used in university workshops and referenced in open-source repos for LLM-based retrieval
  • Bridges the gap between theory and practice for RAG, making advanced LLM techniques accessible to developers
  • Educational Resource: Comprehensive implementation guide for RAG systems
  • Real-world Applications: Demonstrates practical use cases in travel domain

Technical Innovation

  • Advanced RAG Architecture: Combines semantic search with generative AI
  • Context-Aware Responses: Maintains conversation state across interactions
  • Domain-Specific Optimization: Tailored for travel information retrieval
  • Interactive Development: Jupyter notebook format for easy experimentation and learning