CarePulse

• 

  Thumbnail

Inspiration

CarePulse was inspired by the idea that healthcare support should be proactive rather than dependent on patients remembering to check an app. We wanted to explore autonomous AI for daily health tracking.

What it does

CarePulse automatically analyzes patient profiles, conditions, prescriptions, and health information to generate personalized daily health guidance, medication reminders, lifestyle recommendations, and warning indicators.

How we built it

I built CarePulse as an event-driven serverless pipeline using Amazon EventBridge, AWS Lambda, DynamoDB, Bedrock, S3, and SES. EventBridge triggers Lambda, which retrieves patient data, invokes Bedrock, generates the digest, and publishes it.

Challenges we ran into

A major challenge was Amazon Bedrock model invocation. Direct invoke_model testing caused throughput validation errors. We solved this by using the Bedrock Converse API with a Cross-Region Inference Profile.

Accomplishments that we're proud of

I built a headless AI agent that works autonomously without requiring a user to open an application. It can transform stored patient information into personalized content and deliver it through both a web dashboard and email.

What we learned

I learned how powerful generative AI becomes when integrated with event-driven serverless architecture. The project strengthened our understanding of AI orchestration, AWS services, Bedrock integration, automation, and building autonomous workflows.

Built With

• amazon-bedrock
• amazon-dynamodb
• amazon-eventbridge
• amazon-ses
• amazon-web-services
• artificial-intelligence
• aws-lambda
• bedrock
• generative-ai
• healthcare-ai
• lambda
• serverless