AI / ML / 2025
MediCheck
Making uncertainty a little more understandable with probabilistic symptom analysis.
The problem
Symptom information is uncertain. This project explores how an ML application can make probabilistic reasoning understandable through a simple input and analysis workflow.
My role
ML systems, API design, deployment.
A Bayesian ML symptom checker with FastAPI endpoints, a Streamlit interface, and containerized deployment.
The approach
- Structured a Python analysis layer and FastAPI endpoints, with Streamlit providing the interface.
- Used Bayesian machine-learning concepts to translate symptom inputs into probabilistic signals.
- Packaged the application with a multi-stage Docker build, a non-root runtime, and GitHub Actions workflows.
The toolkit
The outcome
Connected symptom input, probabilistic analysis, and an explainable result in a deployable ML application.
An educational engineering project, not a clinical diagnostic service. No clinical validation or patient outcomes are claimed.
Keep exploring