Ask-EHR
ASK-EHR
Throughout March 1-2 2025, I was in a team of 12 to create a streamlined search directory with AI chatbot features.
Nikunj created the idea, then attracted us into the project.
Then, he split the team into story use cases, backend coding, frontend coding and testing and presentation. I was assigned as vice manager, checking the progress of the programme and the presentation.
In the end, we made a working prototype, that I taught them all HuggingFace and a London person taught me Streamlit, and this was the results.
Ask-EHR has the patient’s overview, the doctor’s notes, medical, blood test, imaging, and echo reports.
Then you can type in any questions based on these information:
e.g. “What medicine can he have instead?”
Answer: “If Mr. Smith experiences intolerable side effects with Lamotrigine, an alternative medication for his temporal lobe epilepsy could be Topiramate or Zonisamide, as recommended by the British Epilepsy Association’s guidelines for the management of epilepsy in adults. However, the decision to switch medications should be made in consultation with his neurologist.”
After that I set up the cloud drive where word documents and the presentation can be attached in.
After that, we presented, and then the judges voted us runners-up and got t-shirts.
I was so pleased with Nikunj, and the 10 others for the best work, from the logo, to the work flow, test cases, and learning packages needed for the Python Power application.
Try it at the title: or download here: https://github.com/weepotty/ask-ehr
The Framework and Problem/Solutions
Thank you all, especially you Shona (weepotty), and Nikunj