AI Agent (LLamaIndex + Ollama)

This repository contains my research and development (R&D) code, experiments, and practice projects using LlamaIndex as the core framework. The goal is to explore various AI-driven applications, optimize workflows, and refine my understanding of large language models (LLMs) in different scenarios.

🚀 Tech Stack

• LlamaIndex - Framework for building LLM-powered applications
• Ollama - Local LLM API for efficient model execution
• Qwen2.5 - Selected models: 0.5B and 1.5B

Setup

Install the dependencies

python3 -m venv venv
source venv/bin/activate  # On macOS/Linux
venv\Scripts\activate    # On Windows

Scripts

Human in the loop (HITL) Workflow

File	Description
01_basic_workflow.py	A basic workflow with human-in-the-loop, where workflow take human feedback on response generation
02_agentic_workflow.py	Using agent with wikipedia tool, so that agent can retrieve information from wikipedia. Based on the response, human can provide feedback
03_authentic_api_workflow	Demonstrate how AI agent can access authentic resource taking access from human (credential).