LifeGuard AI: Cardiac Crisis Coordinator

• 

  Demo Interface

The primary inspiration behind the Life-Guard AI Agent was the critical nature of cardiac and stroke emergencies and the need for a swift, reliable coordination system for elderly individuals. The project directly addresses the human latency involved in traditional emergency response by ensuring instantaneous triage and data transfer to a designated caregiver.

The system's core function is AI-driven emergency coordination. It successfully takes a text symptom report (simulating voice input), uses Amazon Bedrock to instantly classify it as a critical event (e.g., stroke), retrieves the patient's pre-registered medical data from DynamoDB, and triggers a guaranteed email alert via Amazon SES to the designated caregiver, making it a complete, automated triage and alerting workflow.

The system was built using a serverless AI Agent architecture. AWS Lambda was utilized as the core orchestrator. A custom Python script manages the workflow: calling Bedrock for LLM reasoning, querying DynamoDB for patient data, and executing the SES alert action. The frontend interface uses native Web Speech API in JavaScript to provide a hands-free input method. The primary challenge run into was persistent network configuration complexity, specifically integrating the local frontend with the Lambda Function URL due to strict CORS security protocols.

Built With

• amazon-bedrock-(claude-3-sonnet)
• amazon-dynamodb
• amazon-dynamodb-(patient-data)
• amazon-ses
• and-amazon-ses-(guaranteed-email-alerting).-the-code-is-primarily-written-in-python-3.11+-(using-the-boto3-sdk)
• and-javascript
• bedrock
• css
• html
• javascript
• lambda
• python
• with-the-frontend-interface-relying-on-standard-html