Boses Berde

sample chat

Inspiration

Building on the UNICEF Green Rising initiative, Project Boses Berde (Filipino for Green Voice) seeks to empower the youth of the Philippines by introducing them to career and training opportunities centered on sustainability and active climate action. The project envisions helping young Filipinos discover meaningful paths that contribute to a greener and more resilient future.

What it does

The core of the application is an AI-powered recommender agent that matches user-provided information—such as age, location, and interests—with a curated database of green jobs and training programs. For the initial release, the system features a static frontend chat and a static backend database. Future iterations will focus on enhancing user management, expanding the database, and integrating the chat interface with existing job boards and training platforms to create a more dynamic, interactive experience.

How we built it

We began with a base chat implementation from existing AWS sample projects and customized it to meet the objectives of Boses Berde. Along the way, we contributed improvements back to the source samples, aligning our work with open collaboration and shared innovation.

Challenges we ran into

One of the major challenges was experimenting with tools that were new to us—such as Nova Act. While this exploration deepened our understanding of the AWS deployment process, it did not lead to a fully functional output within the hackathon deadline.

Accomplishments we’re proud of

Developed and deployed MCP servers dedicated to green jobs and training, integrating them with the main application.
Integrated an Amazon Bedrock Knowledge Base into the app to enhance the AI recommendation layer.
Utilized AWS CDK and CloudFormation for structured and automated deployment of the solution.

What we learned

We gained valuable experience in creating, modifying, and managing AWS AI-related resources, as well as orchestrating cloud infrastructure to support an AI-driven application.