Edugen agent

• 

  Architectural Diagram

Inspiration

The inspiration came from the growing need for accessible, personalized educational tools in a world where students often struggle with access to quality practice materials. As AI becomes more integrated into education, we saw an opportunity to leverage generative AI to create novel, topic-specific questions and answers on demand. Drawing from real-world challenges like exam preparation.

What it does

It is an educational platform that generates novel questions and answers for user-specified topics at customizable difficulty levels.

How we built it

We built it using a serverless AWS architecture for scalability and speed: Frontend: HTML/CSS/JS hosted on S3 with static website hosting. Backend: AWS Lambda handles logic, invoking Bedrock for AI generation. API Layer: API Gateway exposes the endpoint with CORS enabled. Storage: DynamoDB for user questions/scores; S3 for past questions JSON.

Challenges we ran into

Optimizing lambda to run the code: Syntax issues from lambda code formatting caused runtime errors; resolved by cleaning the Python file and redeploying via CLI. CORS and Integration: Frontend-backend mismatches led to "Failed to fetch" errors; fixed by enabling CORS in API Gateway and adding headers. Testing in Serverless: Simulating events required careful JSON payloads; used CloudWatch logs extensively. Choosing a model: deciding which one would be perfect for the task Syncing knowledge base

Accomplishments that we're proud of

Generating high-quality, novel questions with, fully serverless stack on AWS. Overcoming hackathon hurdles: Fixed critical bugs (syntax, CORS) under time pressure, resulting in a demo-ready app. Being able to build a full AI agent

What we learned

How to utilize bedrock - usage, model, agents and knowledge base

What's next for Edugen agent

Authentication: Integrate Cognito for personalized profiles, tracking progress across sessions. Features: multi-question quizzes with scoring and explanations using Bedrock. Expand: Support multi-language generation.

Built With

• amazon-dynamodb
• bedrock
• lambda
• s3