Melli-Mate

• 

  Logo

Inspiration

As a doctor witnessing the devastating impact of poorly managed diabetes, I was determined to create a truly personalized companion called Melli-Mate (from "mellitus" meaning "sweet"). Recognizing the gap in patient support, I envisioned an app that empowers patients, reinforces healthy habits, and ultimately improves outcomes through an iterative patient journey. This innovation aims to ease the burden on our healthcare system.

What Melli-Mate Does

Melli-Mate tackles the challenges of diabetes management with a personalized, empowering approach. Users provide essential details about their health and specific diabetes-related struggles, such as controlling cravings, finding motivation to exercise, or achieving weight loss. Melli-Mate then generates a unique plan addressing their chosen problem, offering motivational guidance on lifestyle adjustments, medication management, and strategies to overcome their specific challenge.

To maximize personalization, Melli-Mate asks follow-up questions throughout the user's journey. This dynamic interaction ensures that the guidance provided evolves with the user’s needs and preferences, making it a truly personalized health companion.

How We Built It

I began by outlining Melli-Mate's core flow. Guided by examples from PartyRock (a programming framework/tool), I iteratively developed the app's conversational structure, prioritizing safety throughout. A key turning point was the realization of the power of focusing on the user's specific struggle. This fuelled Melli-Mate's personalized plans and its dynamic, follow-up questions like "Do you find mornings or evenings work better for exercise?"

Careful tuning of AI parameters, such as temperature and top-p, helped achieve the safe and supportive tone essential for a health companion. This approach to development ensured that Melli-Mate not only addresses users’ health challenges but does so in a manner that is both engaging and comforting.

Challenges We Ran Into

As a doctor with no prior software development experience, I faced a steep learning curve with PartyRock and the overall app-building process. Linking user inputs to the AI text generator was particularly challenging. This initially slowed progress, but dedicated study of input-based PartyRock examples helped me grasp the logic. Overcoming this hurdle significantly boosted my confidence and brought Melli-Mate's dynamic flow within reach.

My initial ambition for a multi-condition healthcare assistant proved too broad, hindering personalized advice. Pivoting solely to diabetes was crucial, allowing me to delve into reliable, condition-specific resources and enabling the creation of Melli-Mate's tailored plans. These challenges underscored the importance of both technical mastery and a deep understanding of patient needs for successful healthcare AI development.

Accomplishments We're Proud Of

Despite having no prior coding experience, building a functional healthcare AI app demonstrates my enthusiasm and dedication. Safety was prioritized in Melli-Mate's design, reflecting my commitment to responsible AI use in healthcare. Melli-Mate's dynamic, iterative question system, which I developed, empowers users to take an active role in managing their health.

Honing my time-management and prioritization skills enabled me to successfully balance the Hackathon with my demanding career as a doctor. Melli-Mate was designed with the potential to improve patient outcomes through personalized support and adherence. This journey from concept to creation has been incredibly rewarding, showcasing the tangible impact that dedicated individuals can have on healthcare innovation.

What We Learned

This journey has expanded my understanding of AI models, their strengths, and their limitations. Taking the "ChatGPT Prompt Engineering for Developers" course significantly improved my skills, and I'm eager to continue learning through further courses. Beyond the technical aspects, I gained a newfound appreciation for AI's potential to transform healthcare, particularly in empowering patients.

This aligns with my vision of Melli-Mate as a tool supporting the shift from paternalistic medicine to truly patient-centered care. I envision AI becoming deeply integrated into patient care and analysis of medical results, potentially leading to significantly improved outcomes. However, this also highlighted the crucial need for strict regulations and responsible use of technology in such a sensitive field. To manage the project alongside my demanding career, I mastered breaking tasks into manageable chunks and using the Pomodoro Technique for focused work sessions. This experience honed my time-management skills, proving invaluable for completing Melli-Mate within the Hackathon constraints.

What's Next for Melli-Mate

Melli-Mate has the potential to become a true companion for people with diabetes, seamlessly integrating into their daily lives and empowering them to manage their health proactively. To enhance user experience and reduce data entry overload, integration with health devices like smartwatches and blood sugar monitors is essential. For example, if blood sugar readings indicate a potential issue, Melli-Mate could provide immediate guidance or prompt the user to log a specific meal for more tailored advice.

Additionally, connecting Melli-Mate to a patient's calendar or reminder apps would support action plan implementation. Further refining the interactive process with reminders could significantly boost user engagement. Recognizing my limitations in taking Melli-Mate to the next level, I'm eager to collaborate with developers who share my vision. My clinical background offers a unique understanding of patient needs, which, when paired with strong technical skills, could make this app even more impactful. My immediate goal is to pilot Melli-Mate with a local diabetes patient group, gathering feedback to guide further iterations and refinements.

Built With

• partyrock