Challenge 1: MLB Search App using Vertex AI Agent Builder

To get familiar with more Google Cloud tools, I chose to build a search app using one of the datasets from the MLB Hackathon repo and Google Cloud's AI Agent Builder.

Comment

Inspiration

The Getting Started video served as inspiration for the web app's functionality and final product.

What It Does

The MLB Search App allows users to ask questions about baseball, MLB Spring Leagues, and the teams within those leagues.

How We Built It

This project was built using:

  • GCloud CLI
  • Gemini
  • Vertex AI Agent Builder
  • Cloud Storage Bucket
  • MLB Hackathon datasets
  • Firebase
  • React
  • TypeScript

Challenges We Faced

Initially, importing datasets from the MLB Hackathon storage bucket using the CLI was challenging. Through trial and error with Gemini and the correct GCloud CLI commands, I discovered a solution. Instead of attempting to import all datasets at once, I imported them individually, which streamlined the process.

Accomplishments We're Proud Of

Successfully integrating the search widget into the web app using an HTML script, environment variables, and Firebase hosting was a major milestone. While deployment into production took time, I managed to start and complete the entire project within a 20-hour window.

What We Learned

I learned how to implement search functionality using Vertex AI and Gemini. Although I did not use DialogFlow in the final deliverable, I explored its capabilities and see significant potential for building engaging conversational experiences.

What's Next for Challenge 1: MLB Search App Using Vertex AI Agent Builder

While I have no immediate plans for further development, I intend to continue exploring AI Agent Builder and Vertex AI for future projects.

Share this project:

Updates