Inspiration
The motivation behind CareBot MediMate (CBMM) stems from the growing pressure on global healthcare systems. Many patients face long wait times for initial triage and post-clinic medication. The situation is more dire for individuals with limited access to the healthcare resources, or individuals for which healthcare is unaffordable. My goal is to build a product that can solve this issue.
What it does
CBMM is an AI-powered healthcare dual-assistant:
- AI Triage Bot: Listens to symptoms and determines triage urgency.
- Virtual Health Assistant: Acts as a medication coach and tracks side effects.
How we built it
- Brain: IBM WatsonX.ai.
- Orchestration: Railtracks framework.
- Data: IBM Db2 and RAG with semantic search.
- Frontend: React ## Challenges we ran into
- Agent Cancellation and task synchronization.
- LLM prompts not generating desired results, causing constant re-prompting for the agent.
Accomplishments that we're proud of
- Eventually building a project that can inform and empower individuals about their health conditions.
What we learned
- Manipulating the prompts for the agents remains a current problem in implementing AI solutions.
- Hallucinations in medical AI can result in severe consequences, ie. delayed treatment, misdiagnose
What's next for GenAI Genesis
- Prompts with optimal results for solving tasks
Built With
- db2
- fastapi
- javascript
- numpy
- pydantic
- python
- railtracks
- react
- watsonx.ai
Log in or sign up for Devpost to join the conversation.