AI Symptom Analysis

Project Overview

Developed during Pandahacks 2025, this AI-powered symptom recognition tool helps users understand potential medical conditions by analyzing freeform symptom descriptions using a hosted LLM.

We currently have a functional MVP that successfully performs real-time symptom recognition and generates condition suggestions.

Approach

Our focus was to build an intuitive, responsive system that could function as a first-touch health support tool, especially for users without access to immediate clinical care. We focused on backend infrastructure, API integration, and prompt engineering to ensure the system was responsive, relevant, and production-aware.

Ongoing Development

We are continuing to develop and refine the application post-hackathon, focusing on the following:

Enhanced Input Filtering & Preprocessing : Implementing logic to clean and normalize inputs

Improved Prompt Structuring : Fine-tuning prompts for edge cases and complex inputs to ensure accuracy and consistency

Frontend Redesign : Upgrading the MVP UI to a more intuitive interface with real-time feedback and successful state handling

Technologies/Skills Used:

Python, Hugging Face Interface, Prompt Engineering, StreamLit