The problem MedX solves

Our solution solves 3 problems we identified:

1. Over time, medical history becomes more and more complicated, especially in case of patients with chronic diseases morever the doctor to patient ratio is declining rapidly and doctors are not able to read the full medical histories and they might miss some important details.
   sol. We intend to solve it by fine-tuning a LLM to summarize the medical histories of patients into desirable length/format.
2. Not enough Attention to allergens can be fatal, sometimes there are some salts that might contain an allergen the patient is allergic to, and in case of such compsumption, it may severly affect the patient. To solve this we Tried to use Web Scrapping to find out common allergens and the medicines which contained them.
3. in Metro and Tier 1 cities, it is common for residents to visit a small-medium private clinic for diagnosis and medical checkup but a common problem observed there is that there often be a long waiting line, so people have to wait at the clinic, to solve this we intent to build a smart appointment system in which when you book an appointment, you will be given a number which corresponds to your priority in the queue as well as it will calculate the time at which your turn would come (approximately).

Challenges we ran into

There were a lot of Bugs, but the most interesting one was when we were web scrapping to collect a dataset to fine tune our model on, pages with "login" got downloaded due to some protocol preventing scrapping from pages. we somehow were able to patch it to some extent that only every 2 in 5 papers were bad which was alright to train the model.