
The project revolved around AI Agents and aimed to create a user-friendly platform for parents to easily find child-related services and activities, from doctors and preschools to festivals and events. Initially, the focus was on showcasing places offering child-centered activities, but later expanded to include events as well. AI Agent was a big part of this project as it serves as the means to an end for interpreting and responding to user queries with the proper answer. The idea behind it was for the AI Agent to quickly and effectively point parents into the right direction, giving them a sense of all the available options in their area.
The main challenge of the project was the tight deadline. The client required rapid development with no room for delays. At the same time, the platform needed to be cost-effective and highly scalable from the start, as it was meant to support a potentially large number of users searching for various types of services and events. A key complexity was building a smart system capable of interpreting user queries and understanding whether the user was looking for a place or an event. This involved integrating real-time location data from the user’s browser and building an AI assistant that could respond intelligently via a chat interface. The platform also had to be simple enough to build quickly, without relying on heavy frontend frameworks or complex infrastructure.
To meet the deadline and build a lightweight yet powerful solution, a serverless architecture was implemented using AWS Lambda, ensuring fast deployment without the need for maintaining servers. All traffic was managed through AWS API Gateway.The core of the app was the Agent AI, an intelligent assistant that interpreted user queries and decided whether the user was looking for a place or an event. The manner in which it was supposed to work was: If the agent detected a query related to a location (e.g., preschools, doctors), it accessed the user’s location via their browser and sent a request to the Google Maps API. If the query was about an event (e.g., family festivals), it pulled results from a vector database designed to handle event data efficiently.Technologies used included Langchain and LangGraph, which powered the AI agent’s logic and dialogue flow. To keep things simple and budget-friendly, no frontend framework was used and the focus was on a minimalistic build with fast iteration.