Inspiration
Healthcare access is often delayed not because facilities don't exist, but because patients and coordinators struggle to find the right facility for their specific needs. Healthcare data is fragmented, inconsistent, and filled with unverified claims. We wanted to build a solution that helps users discover the most appropriate healthcare facility while providing transparent evidence behind every recommendation.
What it does
Raha Care AI is an evidence-based hospital referral copilot built on Databricks. Users can search for a care need such as emergency care, dialysis, maternity services, oncology, or critical care. The platform analyzes healthcare facility records and returns recommended facilities along with location and contact of that facility.
How we built it
We built Raha Care AI using Databricks Free Edition, Unity Catalog, Databricks Apps, SQL, and Python.
Our workflow includes:
- Data ingestion and profiling of healthcare facility records
- Capability and specialty extraction from facility data
- Interactive Databricks App for search and recommendations
- User workflow for reviewing and saving healthcare referrals
Challenges we ran into
The healthcare facility dataset contained inconsistent, noisy, and incomplete information. Many facilities had varying levels of evidence supporting their claimed capabilities. Building a reliable recommendation workflow while communicating uncertainty honestly was one of our biggest challenges.
We also faced challenges around data quality, capability mapping, app deployment, and creating a user experience that remains simple for non-technical healthcare planners and coordinators.
Accomplishments that we're proud of
- Built a working healthcare referral copilot on Databricks
- Designed a workflow suitable for patients, NGOs, and healthcare planners
- Delivered a scalable foundation that can be expanded across healthcare networks
What we learned
This project reinforced the importance of trustworthy AI in healthcare. We learned that transparency, evidence, and uncertainty communication are just as important as the recommendation itself. We also gained hands-on experience building end-to-end applications using Databricks Apps and modern data engineering workflows.
What's next for Raha Care AI
Our vision is to evolve Raha Care AI into a nationwide healthcare navigation platform. Future enhancements include multilingual support, geospatial hospital search, real-time healthcare data integration, personalized referral pathways, and AI-powered care coordination to help millions of people access the right care faster and more confidently.
Log in or sign up for Devpost to join the conversation.