Multi-Agent System for HR and Employee Management

This project is a multi-agent system designed to assist with HR-related tasks and employee management. It leverages the Google Agent Development Kit (ADK) to create a network of specialized agents that work together to provide information and perform actions.

Table of Contents

• Overview
• Architecture
  • Root Agent
  • Employees Agent
  • Google Search Agent
  • Insights Agent
• Getting Started
  • Prerequisites
  • Installation
• Usage

Overview

The system is designed to be a helpful assistant for HR and employee-related queries. It can answer questions about employees, provide HR insights, and perform general web searches for information that it cannot find in its internal knowledge bases.

Architecture

The system is built around a root_agent that acts as a dispatcher, routing user queries to the appropriate specialized agent.

Root Agent

The root_agent is the main entry point for all queries. It is responsible for:

• Understanding the user's intent.
• Deferring to the employees_agent for any questions about employees.
• Deferring to the insights_agent for any questions about HR insights.
• Using the google_search_agent as a fallback for any other questions.

Employees Agent

The employees_agent is responsible for all employee-related tasks. It has the following capabilities:

• Querying Employee Data: It can query a BigQuery database to retrieve employee information. It can perform fuzzy searches and handle variations in names.
• Clocking In/Out: It can clock employees in and out.

Google Search Agent

The google_search_agent is a simple agent that can perform a Google search. It is used as a fallback by the root_agent when a query cannot be answered by any of the other agents.

Insights Agent

The insights_agent is responsible for providing HR insights. It uses Vertex AI Search to search a data store of HR-related documents and information.

Getting Started

Prerequisites

• Python 3.7+
• Google Cloud SDK
• Access to a Google Cloud project with the following APIs enabled:
  • BigQuery API
  • Vertex AI API

Installation

1. Clone the repository:

   git clone https://github.com/NucleusEngineering/hr-agent.git

2. Install the dependencies:

   pip install -r requirements.txt

3. Configure the environment variables:

   Create a .env file (from the dotenv file) and add the following variables:

   GOOGLE_CLOUD_PROJECT="your-gcp-project-id"
   BIGQUERY_DATASET="your-bigquery-dataset"
   BIGQUERY_TABLE="your-bigquery-table"
   VERTEX_AI_SEARCH_DATA_STORE_ID="your-vertex-ai-search-data-store-id"
   

Usage

To run the application, you can use the mystery-runner/main.py script. This will start a local server that you can interact with.

Running the Application

1. Navigate to the mystery-runner directory:

   cd mystery-runner

2. Run the application:

   python main.py

   You should see output indicating that the runner has started and a session has been created. You can then start chatting with the agent in your terminal.