Travel advisor agent using Google Adk
Highly skilled Data Test Automation professional with over 10 years of experience in data quality assurance and software testing. Proven ability to design, execute, and automate testing across the entire SDLC (Software Development Life Cycle) utilizing Agile and Waterfall methodologies. Expertise in End-to-End DWBI project testing and experience working in GCP, AWS, and Azure cloud environments. Proficient in SQL and Python scripting for data test automation.
Travel Advisor Agent powered by Google ADK and Gemini, capable of answering travel questions and acting like a personal AI travel consultant.
What is Google ADK (Agent Developer Kit)?
Google ADK is a framework designed to help developers build autonomous, tool‑using AI agents. It provides:
Agent lifecycle management
Tool integration (APIs, functions, external services)
Memory & context handling
Seamless integration with Gemini models
In short, ADK gives you the scaffolding to build intelligent agents without reinventing the wheel.
The Architecture: How the Travel Advisor Works
The goal of this project was to create an agent that doesn't just "chat" about travel but actually performs tasks. The core architecture consists of:
The Brain (Gemini 2.0 Flash lite): Processes user intent and decides which tools to call.
The Framework (Google ADK): Orchestrates the communication between the model and the environment.
The Tools: Custom Python functions that the agent can execute to fetch the required information.
Project Structure
Github repository contains the following key files:
| File | Purpose |
agent.py | Core logic for the travel advisor agent |
__init__.py | Package initialization |
.env | Environment variables (API keys, etc.) |
The heart of the project is agent.py, where the agent is defined and configured.
How the Travel Advisor Agent Works
At a high level, the agent:
Loads the Gemini model
Accepts user travel queries
Uses ADK to manage conversation context
Generates travel recommendations or answers
Returns structured, helpful responses
Here’s a conceptual flow:
User Query → ADK Agent → Gemini Model → Travel Recommendation → Response
Execution Output
GitHub Repo
https://github.com/puthanvipin/travel-advisor-agent-using-google-adk
