Inspiration
Every year, thousands of people are newly diagnosed with HIV — and many of them face the diagnosis alone, with fear, stigma, and little emotional or medical guidance. Searching online often leads to unreliable answers: tools like ChatGPT can “hallucinate” or provide information from non-official sources, which only increases confusion and anxiety. We wanted to change that by building a trusted, compassionate digital companion.
What it does
Contigo.care is an AI-powered agent that supports patients in the first critical month after an HIV diagnosis.
- Provides reliable, verified medical information instead of misinformation.
- Offers emotional support, detecting signs of anxiety or depression.
- Sends personalized reminders for medication and appointments.
- Connects patients securely with psychologists and doctors when needed. It’s like a Buddy + Copilot for health, combining empathy, accuracy, and professional oversight.
How we built it
We used AWS to design an AI agent that is both functional and scalable:
- Amazon Bedrock for conversational AI (ContigoBot).
- Amazon Comprehend Medical to detect emotional and clinical signals in text.
- AWS Lambda + DynamoDB for workflow orchestration and patient state management.
- Amazon SNS / Pinpoint for notifications and reminders.
- Amazon Cognito for secure authentication and GDPR-compliant data protection.
- The frontend is a lightweight web/mobile app where patients interact with the agent.
Challenges we ran into
- Ensuring empathy + clinical accuracy in the AI’s responses.
- Designing with data privacy and security at the center (sensitive health data).
- Creating a workflow that complements doctors instead of replacing them.
- Reducing the vision of a full digital health ecosystem into an MVP functional enough for this hackathon.
Accomplishments that we're proud of
- Built an MVP of an AI agent that does not hallucinate, grounded in verified data sources.
- Designed a solution that empowers patients while reducing hospital burden.
- Created a prototype that is technically feasible and socially impactful.
- Showed how AI can make the healthcare system more humane, not less.
What we learned
- Patients need trustworthy AI companions, not generic chatbots.
- Emotional health is as critical as physical health in treatment adherence.
- Building with AWS makes it possible to combine AI, data, and security at scale.
- A strong storytelling + technical design is key to convincing both hospitals and judges.
What's next for Contigo.care
- Pilot at Hospital La Paz (Madrid) with 100 patients to measure real-world impact.
- Expand across Spanish hospitals and integrate with national healthcare systems.
- Add modules for family education, prevention, and sexual health awareness.
- Scale into a national digital ecosystem for HIV, combining patients, professionals, and institutions.
Built With
- amazon-bedrock
- amazon-comprehend-medical
- amazon-dynamodb
- amazon-sns/pinpoint
- amazoncognito
- aws-lambda
- dynamodb
- figma
- github
- javascript
- kiro
- python
- quicksight
- react
Log in or sign up for Devpost to join the conversation.