Building a Multi-Agent System to Find My Next Neighborhood Abroad

{ "author": { "bio": null, "image": null, "name": "Caroline Scholles" }, "body": [ { "_key": "fdfba4ff4478", "_type": "block", "children": [ { "_key": "c73ef6c6611f", "_type": "span", "marks": [], "text": "This year I'm moving to Portugal." } ], "markDefs": [], "style": "normal" }, { "_key": "a6213d05d3e0", "_type": "block", "children": [ { "_key": "1be2f85b4c07", "_type": "span", "marks": [], "text": "I'm not sure if it will be permanent. I've moved before with every intention of staying forever, but working remotely gives me so much flexibility that it's hard to commit to one place indefinitely." } ], "markDefs": [], "style": "normal" }, { "_key": "0e5a9d935cd2", "_type": "block", "children": [ { "_key": "94640e769057", "_type": "span", "marks": [], "text": "After spending countless hours manually searching for rental properties across multiple listing sites, I decided to automate part of the decision-making. I built a home-hunter-agent using Google's Agent Development Kit (ADK) that takes a target location and my preferences as input, then returns ranked recommendations for the best neighborhoods." } ], "markDefs": [], "style": "normal" }, { "_key": "ff3ea76a5964", "_type": "block", "children": [ { "_key": "4a557cf09cc5", "_type": "span", "marks": [], "text": "This is my first time implementing a multi-agent architecture. The home-hunter-agent coordinates multiple specialized sub-agents, each handling different aspects of the search: safety analysis, walkability scores, and proximity to amenities via the Google Maps API. It generates a livability recommendation report that I reference while searching for rentals." } ], "markDefs": [], "style": "normal" }, { "_key": "60ec4bd0edab", "_type": "block", "children": [ { "_key": "1b5444113df6", "_type": "span", "marks": [], "text": "Given a target location and your lifestyle preferences, the home-hunter-agent automatically:" } ], "markDefs": [], "style": "normal" }, { "_key": "b60386329f36", "_type": "block", "children": [ { "_key": "a24de59fd70a", "_type": "span", "marks": [], "text": "Analyzes neighborhood safety and walkability" } ], "level": 1, "listItem": "bullet", "markDefs": [], "style": "normal" }, { "_key": "91b9ab4206fe", "_type": "block", "children": [ { "_key": "fdb35b0b3033", "_type": "span", "marks": [], "text": "Maps nearby amenities using the Google Maps Places API (cafes, gyms, parks, restaurants, etc.)" } ], "level": 1, "listItem": "bullet", "markDefs": [], "style": "normal" }, { "_key": "50e1343c9d5e", "_type": "block", "children": [ { "_key": "65ed9c650c97", "_type": "span", "marks": [], "text": "Calculates proximity scores for places that matter to you" } ], "level": 1, "listItem": "bullet", "markDefs": [], "style": "normal" }, { "_key": "8a4cd3682523", "_type": "block", "children": [ { "_key": "abd43970f66b", "_type": "span", "marks": [], "text": "Generates livability recommendations based on your priorities" } ], "level": 1, "listItem": "bullet", "markDefs": [], "style": "normal" }, { "_key": "c2f4974c7253", "_type": "block", "children": [ { "_key": "0b8e70d3e518", "_type": "span", "marks": [], "text": "Produces a comprehensive neighborhood analysis report" } ], "level": 1, "listItem": "bullet", "markDefs": [], "style": "normal" }, { "_key": "0a07bfb5a7ca", "_type": "block", "children": [ { "_key": "fe4330061134", "_type": "span", "marks": [], "text": "The agent doesn't drill down to individual properties—I just wanted to identify which neighborhoods fit my lifestyle best. You can customize it with parameters like budget and anything else you'd search for on Google Maps (which, when you think about it, is incredibly powerful)." } ], "markDefs": [], "style": "normal" }, { "_key": "4925d6e10f18", "_type": "block", "children": [ { "_key": "755fff455067", "_type": "span", "marks": [], "text": "It's already surfaced some interesting neighborhoods I hadn't considered, which I'm now sharing with my relocation consultant. (Yes, I'm definitely not relying completely on a machine to decide where I'll live for the next few months.)" } ], "markDefs": [], "style": "normal" }, { "_key": "9be494b5b3a9", "_type": "block", "children": [ { "_key": "d8ce95870c86", "_type": "span", "marks": [ "strong" ], "text": "GitHub:" }, { "_key": "92a85ad544c0", "_type": "span", "marks": [], "text": " " }, { "_key": "c94c9ed2cc8b", "_type": "span", "marks": [ "4d49bcac0cdf" ], "text": "https://github.com/carolinescholles/home-hunter-agent" } ], "markDefs": [ { "_key": "4d49bcac0cdf", "_type": "link", "href": "https://github.com/carolinescholles/home-hunter-agent" } ], "style": "normal" }, { "_key": "acb0b9953535", "_type": "block", "children": [ { "_key": "9e7348296318", "_type": "span", "marks": [], "text": "" } ], "markDefs": [], "style": "normal" }, { "_key": "90347a2e6a14", "_type": "block", "children": [ { "_key": "4486f6a8ff54", "_type": "span", "marks": [ "strong" ], "text": "If you want to try this:" } ], "markDefs": [], "style": "normal" }, { "_key": "786440ef4c14", "_type": "block", "children": [ { "_key": "d7bcc57bd666", "_type": "span", "marks": [], "text": "Follow lessons 1-7 from " }, { "_key": "48242e19d462", "_type": "span", "marks": [ "5d11223e4c55" ], "text": "Google's Advent of Agents" } ], "level": 1, "listItem": "bullet", "markDefs": [ { "_key": "5d11223e4c55", "_type": "link", "href": "https://adventofagents.com/" } ], "style": "normal" }, { "_key": "88ec2be771d5", "_type": "block", "children": [ { "_key": "d6aa297fb056", "_type": "span", "marks": [], "text": "Check out the " }, { "_key": "4328a99ffbc8", "_type": "span", "marks": [ "bba0cf1ab1fd" ], "text": "ADK samples repo" }, { "_key": "7387f1b62e3b", "_type": "span", "marks": [], "text": " and test how to use the library yourself" } ], "level": 1, "listItem": "bullet", "markDefs": [ { "_key": "bba0cf1ab1fd", "_type": "link", "href": "https://github.com/google/adk-samples" } ], "style": "normal" }, { "_key": "3f2c1f610252", "_type": "block", "children": [ { "_key": "9464d86d06b3", "_type": "span", "marks": [], "text": "You can also use " }, { "_key": "6d3e19329c23", "_type": "span", "marks": [ "15d0429ff1da" ], "text": "my github" }, { "_key": "fc4bc4276141", "_type": "span", "marks": [], "text": " as a boilerplate and start making changes according to your needs using AI or by hand." } ], "level": 1, "listItem": "bullet", "markDefs": [ { "_key": "15d0429ff1da", "_type": "link", "href": "https://github.com/carolinescholles/home-hunter-agent" } ], "style": "normal" } ], "categories": [ { "title": "Tools & Workflows" } ], "faqs": [ { "answer": "The Agent Development Kit (ADK) is a framework from Google that simplifies the process of building AI agents and multi-agent systems. It provides tools for designing, developing, and deploying agents that can perform tasks, interact with other agents, and utilize external tools.", "question": "What is the Agent Development Kit (ADK)?" }, { "answer": "The home-hunter-agent analyzes neighborhoods based on safety, walkability, and proximity to amenities using the Google Maps API. It generates livability recommendations based on customized priorities to help find the best neighborhoods.", "question": "What does the home-hunter-agent do?" }, { "answer": "Yes, the home-hunter-agent can be customized with parameters like budget and other criteria searchable on Google Maps.", "question": "Can the home-hunter-agent be customized?" }, { "answer": "The code is available on GitHub: https://github.com/carolinescholles/home-hunter-agent.", "question": "Where can I find the code for the home-hunter-agent?" } ], "mainImage": null, "publishedAt": "2026-01-02T00:57:00.000Z", "seo": { "title": "Building a Multi-Agent System to Find My Next Neighborhood Abroad" }, "slug": { "_type": "slug", "current": "multi-agent-system" }, "takeaways": [ "The author is moving to Portugal and is leveraging remote work flexibility to explore living abroad.", "A multi-agent system was built using Google's Agent Development Kit (ADK) to automate neighborhood research for relocation.", "The home-hunter-agent uses specialized sub-agents for safety analysis, walkability scores, and proximity to amenities.", "The agent provides a comprehensive neighborhood analysis report to identify suitable neighborhoods based on lifestyle preferences.", "While automation helps, the author emphasizes the importance of consulting with relocation experts and not solely relying on machines." ], "title": "Building a Multi-Agent System to Find My Next Neighborhood Abroad" }

This year I'm moving to Portugal.

I'm not sure if it will be permanent. I've moved before with every intention of staying forever, but working remotely gives me so much flexibility that it's hard to commit to one place indefinitely.

After spending countless hours manually searching for rental properties across multiple listing sites, I decided to automate part of the decision-making. I built a home-hunter-agent using Google's Agent Development Kit (ADK) that takes a target location and my preferences as input, then returns ranked recommendations for the best neighborhoods.

This is my first time implementing a multi-agent architecture. The home-hunter-agent coordinates multiple specialized sub-agents, each handling different aspects of the search: safety analysis, walkability scores, and proximity to amenities via the Google Maps API. It generates a livability recommendation report that I reference while searching for rentals.

Given a target location and your lifestyle preferences, the home-hunter-agent automatically:

Analyzes neighborhood safety and walkability
Maps nearby amenities using the Google Maps Places API (cafes, gyms, parks, restaurants, etc.)
Calculates proximity scores for places that matter to you
Generates livability recommendations based on your priorities
Produces a comprehensive neighborhood analysis report

The agent doesn't drill down to individual properties—I just wanted to identify which neighborhoods fit my lifestyle best. You can customize it with parameters like budget and anything else you'd search for on Google Maps (which, when you think about it, is incredibly powerful).

It's already surfaced some interesting neighborhoods I hadn't considered, which I'm now sharing with my relocation consultant. (Yes, I'm definitely not relying completely on a machine to decide where I'll live for the next few months.)

GitHub: https://github.com/carolinescholles/home-hunter-agent

If you want to try this:

Follow lessons 1-7 from Google's Advent of Agents
Check out the ADK samples repo and test how to use the library yourself
You can also use my github as a boilerplate and start making changes according to your needs using AI or by hand.