README

Description

Develop a blood-pressure Q&A agent to answer the following questions based on patients’ vitals and medication data provided

1. Did the patient have Hypertension/Hypotension given blood-pressure records from vitals?

2. Did the patient get the medication order to treat hypertension/hypotension if any?

Workflow

1. Agent loads data from data folder and saves in DB.
2. Agent then queries OpenAI GPT-4 to with above questions and data retrieved from DB.
3. Agent Responses stored in DB are retrieved and analyze against heuristic logic predictions.

Submission

From the Video Submission and Submission Logs, we can see that patient 1 has Hypertension and has been provided with the appropriate medicine. Patient 2 however has Hypotension has not yet been provided with the appropriate medicine.

Data and DB

Datasets need to be loaded in data\___.csv folder and DB used is MongoDB Atlas (URI in constants\constants.py) Please add a .env file where you can put your OPENAI_KEY. Note - OpenAI doesn't allow pushing these keys on Github.

LLM Agent Templates

All templates are visible click here

Video and Submission Log

Please see the Video/Log Submission where the Agent runs from Docker container, queries LLM and shows responses against heurstic logic

Video click here Logs click here