Inspiration

Inspiration came from Real Estate Agent, who helps at various steps in purchasing a house. Why not have a Health Care agent who helps in getting the best Medical Care.

What it does

The Health Care Agent talks to users and suggests best procedures based on healthvariety (types of procedures, recovery, complications, success rates, costs), number of people visiting the health care facilities(trips) and google businesses locations (medical facility)

How we built it

Using Mosaic AI Agent Framework and llm (Meta Llama 3.3 70B Instruct), combining data from healthvariety (for medical procedures at particular locations), google_maps_businesses(finding only medical businesses), nyctaxi (trips to medical facilities)

Challenges we ran into

Finding out which datasets to use to get the most accurate information to find the best medical facility. Dataset was too huge for memory, hence sampled the data.

Accomplishments that we're proud of

Able to get answers related to the procedures like success rates, complications, recovery time, availability, cost, etc. Able to find the what procedures are performed at a particular location.

What we learned

How to extract datasets from multiple sources. Integrate them into the agent. Registering the agent and running the agent. Prompting the agent and getting the most accurate answers, that are backed up by data.

What's next for health_agent

Smoothen the agent responses and use the feedback to improve accuracy. Using the entire datasets and more fields to improve accuracy. Assigning weights to different datasets to improve accuracy.

Built With

Share this project:

Updates