About the Project

🚀 Inspiration

The idea for this project came from observing how hospitals often struggle with long waiting queues, unorganized appointment systems, and lack of proper patient prioritization. Many patients are unsure which specialist to consult, and in critical situations, delays can have serious consequences. This motivated us to explore how AI could be used to make healthcare access faster, smarter, and more efficient.

🧠 What We Learned

Throughout the development of this project, we gained hands-on experience in building a full-stack application with a modular and scalable architecture. We learned how to design role-based systems for patients, hospitals, and administrators, and how to create intuitive user interfaces using modern frontend technologies. We also explored how large language models (LLMs) can be used for dynamic question generation and symptom-based analysis, along with prompt engineering to control outputs. Additionally, we improved our understanding of state management, routing, and system design.

🛠️ How We Built It

The project was built using a structured approach, starting with frontend development using React (Vite) and Tailwind CSS to create clean and responsive dashboards. The application was divided into three main modules: patient, hospital, and admin, each with its own functionality. For AI capabilities, we used a locally running Gemma LLM via LM Studio, which allowed us to generate dynamic follow-up questions and analyze patient symptoms without relying on external APIs. A backend layer using FastAPI was planned to handle communication between the frontend and the AI model, ensuring smooth data flow and structured responses.

⚙️ Challenges Faced

One of the biggest challenges was designing an AI-like system that could dynamically generate meaningful questions without relying on external APIs or large-scale training. Ensuring that the model outputs remained structured, relevant, and safe (without suggesting treatments or medications) required careful prompt design and testing. Another challenge was maintaining a clean and scalable frontend architecture while working under time constraints. Balancing functionality with simplicity for an MVP was also a key consideration, as we needed to prioritize core features while keeping the system intuitive and efficient.

Overall, this project helped us understand how intelligent systems can be integrated into real-world applications to solve practical problems, especially in critical domains like healthcare.

Built With

• antigravity
• chatgpt
• css
• fastapi
• figma/canva
• gemma
• github
• javascript
• llm
• lm
• lmstudio
• python
• react
• supabase
• tailwind