MedEase - Intelligent Medication Assistant

Inspiration

Our project was inspired by a very common real-world problem: people forgetting to take their medications or supplements correctly. Many individuals — especially elderly users, immigrants with language barriers, students taking supplements, and busy adults — struggle to manage multiple medications with different schedules, food restrictions, and timing requirements.

That idea became MedEase, an intelligent medication scheduling assistant that helps users organize medications, generate optimized daily plans, and simplify medical information using AI.

What it does

MediEase helps users:

Add and manage medications
Generate optimized daily medication schedules
Track medication history and adherence
View a “Today Dashboard” with upcoming doses
Receive medication reminders through Discord webhooks
Use AI simplify mode to explain medications in easier language

One of our key features is the Daily Plan Generator. Users can:

Select medications from their existing medication list
Enter wake-up time, sleep time, and meal times
Let AI generate an optimized medication schedule

The AI attempts to create a schedule that minimizes conflicts and improves convenience for the user.

How we built it

Frontend

React
TypeScript
Tailwind CSS

The application uses a responsive dashboard layout with:

Left sidebar navigation
Main medication timeline dashboard
Contextual AI assistant panel

We focused heavily on creating a clean and accessible user experience with large readable cards and simplified navigation.

Backend

Python
MongoDB
FastAPI
Gemini The backend was designed to support:
Medication management
Schedule generation
User profile storage
Medication history tracking We structured the project around RESTful API principles to keep the architecture modular and scalable.

AI Integration

We integrated the Gemini API to power the schedule generation and simplify mode features.

One major part of development was carefully designing:

the AI prompts
the structure of input data
the expected output format

Before sending data to Gemini, we had to structure user information in a way the model could properly understand:

medication list
dosage information
meal schedules
wake/sleep times
timing constraints

After receiving the AI response, we then had to:

extract only the necessary fields
validate the data
map the response into our defined backend models
store the cleaned data into the database

This process required a lot of iteration because AI responses are not always perfectly structured.

Conceptually, our optimization logic can be thought of as:

[ OptimalSchedule = \arg\min (timing\ conflicts + missed\ dose\ risk + inconvenience) ]

Challenges we ran into

Structuring AI Input and Output One of the hardest challenges was working with AI-generated data.

We had to:

carefully engineer prompts
define structured response formats
extract only valid fields from Gemini responses
convert AI output into backend-ready objects Since AI responses can vary, ensuring reliable and consistent parsing became a major technical challenge.

Balancing Smart Scheduling Logic Medication timing is more complicated than simple reminders.

Different medications may require:

spacing between doses
food restrictions
sleep considerations
multiple daily intervals Generating a schedule that works realistically with a user’s daily routine required balancing many constraints simultaneously.

Designing for Simplicity Another challenge was preventing the app from becoming overwhelming.

At first, the dashboard contained too much information, which increased cognitive load. We redesigned the interface to focus on:

clarity
fast interaction
immediate visibility of the next medication task

We learned that healthcare applications need to prioritize simplicity over feature overload.

Accomplishments that we're proud of

We realized that most reminder apps only send notifications, but they do not actually help users understand when medications should optimally be taken or explain instructions in simple language. We wanted to build something smarter and more human-centered. MediEase is more than a reminder app — it is an intelligent daily health companion designed to simplify medication management and reduce missed doses through accessibility-focused design and AI assistance.

Our goal was to create something practical, human-centered, and genuinely helpful for everyday life.

What we learned

Through this project, we learned:

how to structure AI-integrated applications
the importance of prompt engineering
how to validate and transform AI-generated data
how to design scalable React dashboard layouts
the importance of accessibility-focused UX design

We also learned that building AI-powered applications is not just about connecting an API — it requires strong data modeling, validation, and thoughtful user experience design.

What's next for MedEase

In the future, we would like to add:

Smart refill prediction
Camera-based medication recognition
Wearable device integration
Personalized adherence analytics
More advanced AI health explanations We also want to improve multilingual support to better help immigrant and ESL communities.