GlucoWise

Intro and Inspiration

Diabetes in a major health issue in our country. Being in medical profession, I have been exposed to a lot number of cases including of my grandmother where monitoring becomes a major issue. The next issue they encounter is the adherence to medication and lack of reminder and motivation for lifestyle modifications. Adding to that, the high recurring cost of CGM i.e. Rs.5ooo which runs for max 14 days also becomes a economic liability on the already suffering patient. For the complications, like retinopathy, neuropathy, nephropathy or diabetic foot, these once there can't mostly be reversed and thus early detection, prevention and advice for timely, proper care esp in case of diabetic foot becomes a key player for better prognosis. These very problems prompted me to find solution for them and thus led to the idea of this app.

How Does This Work

GlucoWise is an AI-powered mobile application designed to simplify diabetes management while empowering patients to take control of their health. It begins with a short calibration phase using 7–14 days of CGM (Continuous Glucose Monitoring) data, combined with user inputs such as meals, medications, sleep, and symptoms. This data trains the AI model to understand the user's unique glucose patterns.

After this learning period, the app functions with just 1 daily blood sugar reading or even less depending on patient condition, the clinical recommendation and compliance. The AI continues to predict glucose trends using time-series and regression models, providing proactive alerts before potential spikes or crashes. It also incorporates a feedback mechanism where the app evaluates the accuracy of its predictions and refines itself accordingly, ensuring adaptive learning over time.

GlucoWise further integrates intelligent medication reminders and personalized health prompts, enhancing user adherence and reducing the risk of missed doses. The app minimizes the need for constant monitoring while maintaining accuracy through periodic checks and smart recalibration. Designed with scalability and real-world usability in mind, GlucoWise bridges the gap between high-end monitoring devices and patient-friendly tech—making diabetes care smarter, more predictive, and more personalized.

How we built it

So, for the app part, the prototype has been built using java on Android-Studio which presently has user login, data collection(Meal stats, Medication, Symptoms, Sleep) and storage portal (using sql), data display, graph plotting and basic advices displayed on its dashboard, a CGM pairing and data import panel and a support page. For visualization, demo and getting idea for finetuning of UI we have used Bubbleapps.io a minimal to no code developer. For the AI part, the basic rules and guidelines have been almost laid out using the standard references and various scientific research publications and known parameters and their respective effects on blood glucose curve. Further work on the app and AI is underway.

Challenges we ran into

Major challenge so far was the app development part and lack of prospective data for the AI especially when doing in such a short interval of time. But I guess both the issues will get managed with time and further exploration and collaboration with stakeholders.

Accomplishments that we're proud of

The idea that has stemmed, the dedication that has been put to it is something I am quite proud of. Starting from scratch, just plethora of problems to be able to make a roadmap and short of a prototype in such a short time interval is something incredible for me.

What we learned

No problem is big or difficult to solve if there is passion for it.

What's next for GlucoWise

After complete development of app and initial significant training of AI on available records and data, I lookforward to collaboration with stakeholders i.e. clinicians and patients for their inputs and suggestion towards the product. Then after calibration and testing on few actual patients' live data, I plan to roll out the app in the market. I feel confident that this would prove to be a revolutionary product. I have also planned of few future steps like integration of Image based Diabetic Foot screening aided with the already present glucose levels and monitoring of symptoms if any aiding in the earliest detection of the complication. I am also looking forward to some non invasive spectrometric techniche to determine blood glucose level.. maybe by using some end glycated product like glycated albumin etc or even other modalities with shorter and early change detection.

Built With

• android-studio
• bubble
• java
• sqlite