AI CareCoach – Intelligent Long-Term Care Training Assistant

Empowering caregivers with AI — learn, practice, and grow anytime.

Comment

Inspiration

China’s Long-Term Care Insurance (LTCI) system is rapidly expanding, yet the training quality of community caregivers remains uneven. Most caregivers lack structured learning tools, interactive practice, and real-time feedback. AI CareCoach was inspired by this gap — to create an AI-powered training assistant that combines professional nursing standards with accessible, empathetic technology.

What it does

AI CareCoach is an intelligent long-term care training platform that helps caregivers learn essential nursing skills step by step. It provides structured modules across five domains — daily care, basic nursing, symptom management, functional training, and psychological support — each divided into junior, intermediate, and senior levels that unlock progressively. Learners can: • Study structured “What–Why–When–Scenario” knowledge points • Practice with interactive quizzes and real-time feedback • Review with flashcards that connect reasoning and actions • Chat with XiaoHu, the AI assistant, for contextual explanations and module recommendations • Track learning progress, study time, and unlocked achievements via an interactive dashboard

How we built it

We built AI CareCoach as a serverless web app using: • Frontend: HTML5, CSS3, JavaScript (Vanilla JS) • Cloud Backend: AWS Lambda (Python) + Amazon S3 for content retrieval • Data: JSON-based RAG (Retrieval Augmented Generation) knowledge sets derived from China’s official Long-Term Care Worker Training Textbook • AI Assistant: In-browser reasoning and voice response via Web Speech API • Local Storage: For offline tracking of progress, achievements, and study time The architecture enables both cloud-based scalability and local offline usability, suitable for caregivers in low-bandwidth environments.

Challenges we ran into

• Configuring CORS and IAM policies for secure Lambda–S3 communication
• Structuring real textbook data into modular “What–Why–When–Scenario” formats
• Designing a Coursera-style UI that remains lightweight and fully responsive
• Building an empathetic AI assistant that sounds supportive, not robotic
• Balancing cloud integration with offline accessibility for rural caregivers

Accomplishments that we're proud of

• Developed a fully functional AI training assistant with AWS serverless architecture
• Designed five complete learning modules with 15 courses and three-level progression
• Built XiaoHu, a bilingual AI care assistant with contextual reasoning and voice replies
• Integrated flashcard review, real-time quiz evaluation, and progress visualization
• Produced a bilingual demo and video presentation showcasing both usability and empathy

What we learned

• How to design AI-driven educational tools that translate textbook theory into real caregiving practice
• The power of AWS Lambda + S3 for scalable, serverless learning platforms
• The importance of UX in healthcare education — empathy and clarity matter as much as code
• That even simple localStorage tracking can create strong motivation through visual feedback

What's next for AI CareCoach – Intelligent Long-Term Care Training Assistant

• Integrate Amazon Bedrock for conversational reasoning and dynamic feedback
• Use SageMaker + DynamoDB to analyze skill patterns and generate personalized learning paths
• Expand to multilingual support (EN, JP, KR) for international caregiving programs
• Launch a PWA offline version for low-connectivity communities
• Collaborate with local LTCI providers to implement AI CareCoach in real-world caregiver training

Built With

Share this project:

Updates