Get in Touch
We can help with Generative AI Integrations, Solutions Architecture, DevOps, Healthcare Diagnostics Apps
Send emailFrequently Asked Questions
The heartcheckapp is powered by Amazon Bedrock foundation model (FM) Anthropic Claude 4 Sonnet.
Claude Sonnet 4.0 is guided by a written set of ethical principles, ensuring safe and reliable interactions. This approach emphasizes transparency, helpfulness, and harmlessness.
The heartcheckapp has been designed so that the input (prompts) are strickly controlled to ensure the model's ethical use and accuracy.
System prompts (guardrails) ensure that the model's responses remain within the limited scope of the app's preventative healthcare mandate.
The heartcheckapp is compliant with the evolving landscape of AI regulations, including the General Data Protection Regulation (GDPR) compliance guidelines and the EU AI Act.
The heartcheckapp uses Amazon Comprehend Medical, a HIPAA-eligible natural language processing (NLP) service that uses machine learning, to quickly and accurately extract protected health information (PHI) from medical reports.
The PHI is then anonymized and stored securely and separately from the medical data.
The heartcheckapp will provide treatment recommendations based on the assessed risk level of the user. A reputable and practicing Cardiologist from a leading medical institution will have design input into the system prompts of the model generating these recommendations. The model's recommendations will be audited and reviewed regularly by a Cardiologist.