Academia Agent

Inspiration

Thinking about planning for future children's education options as well as creating something to help family and friends with children inspired this project. The goal was to make research about post-secondary school options simpler and more accessible.

What it does

Academia Agent helps people complete research about post-secondary school options, providing tools and insights to make informed decisions about education pathways.

How we built it

• Frontend: Built using Next.js for a responsive and modern UI.
• Authentication: Leveraged AWS Amplify for Auth after pivoting from custom solutions.
• Infrastructure: Custom AWS resource deployment using AWS CDK.
• AI Agent:
  • Built using AWS Bedrock AgentCore.
  • Utilized AWS Nova Micro model for reasoning and decision-making.
  • Integrated AgentCore Tools for research and planning assistance.

Challenges we ran into

• Frontend Complexity: Working with Next.js and customizing authentication was challenging.
• Auth Customization: Implementing easier methods like third-party federation and passwordless authentication (email OTP, passkeys) required significant effort.
• Onboarding UX: Making sign-up and sign-in seamless was harder than expected.

Accomplishments that we're proud of

• Successfully pivoted to AWS Amplify for Auth, reducing complexity.
• Deployed the frontend using AWS Amplify, saving time on resource configuration and CloudFront setup.

What we learned

• The depth of work required for authentication in web applications.
• How to use Bedrock AgentCore effectively with Strands SDK for AI-driven workflows.

What's next for Academia Agent

• Add more authentication methods for sign-up/sign-in.
• Customize the UI for authentication flows.
• Introduce tools for research and decision drafting.
• Add models/agents for:
  • Generating visual roadmaps.
  • Enabling audio-based interactions for a more natural user experience.

Built With

• agentcore
• amazon-web-services
• amplify
• bedrock
• nextjs
• python
• strands
• typescript