Agentic Systems

A comprehensive implementation of the paper "Agentic Systems: A Guide to Transforming Industries with Vertical AI Agents" by Fouad Bousetouane.

Key Architectures

RAG Agent Router with Domain-Specific Vector Databases

RAG Orchestrated Multi-Agent System for Multi-Domain Knowledge Retrieval

Overview

This repository provides a technical implementation of agentic systems as described in the paper. It showcases the architecture and components of AI agents powered by Large Language Models (LLMs), particularly using the Ollama deepseek-r1:14b model for local deployment.

Architecture

The implementation follows the modular architecture described in the paper, consisting of:

Core Components

• Memory Module: Stores context, history, and domain knowledge
• Reasoning Engine: LLM-powered decision-making core (using deepseek-r1:14b)
• Cognitive Skills: Domain-specific inference capabilities
• Tools: External integrations and API connectors

Agent Types

• Task-Specific Agents: Specialized for discrete tasks (e.g., RAG Router)
• Multi-Agent Systems: Collaborative frameworks for complex tasks
• Human-Augmented Agents: Systems with human oversight and feedback

Getting Started

Prerequisites

• Python 3.8+
• Ollama with deepseek-r1:14b installed

Installation

1. Clone the repository:

git clone https://github.com/yourusername/agentic-systems.git
cd agentic-systems

2. Install the required dependencies:

pip install -r requirements.txt

3. Install Ollama and pull the deepseek-r1:14b model:

# Install Ollama from https://ollama.ai/
# Then pull the model:
ollama pull deepseek-r1:14b

4. Install the package in development mode:

pip install -e .

Running Examples

The repository includes three example implementations:

1. Customer Support (Human-in-the-Loop Agent):

python examples/customer_support/hitl_customer_agent.py

2. Knowledge Management (RAG Router Agent):

python examples/knowledge_management/rag_router_demo.py

3. Legal Analysis (Multi-Agent System):

python examples/legal_analysis/multi_agent_legal_system.py

See the examples directory for more details on each implementation.

Project Structure

agentic_systems/
├── agents/                # Agent implementations
│   ├── task_specific/     # Task-specific agents
│   ├── multi_agent/       # Multi-agent systems
│   └── human_augmented/   # Human-augmented agents
├── core/                  # Core components
│   ├── memory/            # Memory implementations
│   ├── reasoning/         # Reasoning engines
│   ├── cognitive_skills/  # Specialized inference capabilities
│   └── tools/             # External integrations
└── examples/              # Example implementations
    ├── customer_support/  # HITL agent for customer support
    ├── knowledge_management/ # RAG Router for knowledge retrieval
    └── legal_analysis/    # Multi-agent system for legal analysis

Documentation

Detailed documentation for each component is available in their respective directories.

License

MIT

Citation

If you use this implementation in your work, please cite the original paper:

Bousetouane, F. (2024). Agentic Systems: A Guide to Transforming Industries with Vertical AI Agents.