CTAID POC
Portfolio demonstration
A CT atherosclerosis detection research app, rebuilt as a clean proof of concept.
CTAID POC is an English-only portfolio demo. It shows how a small machine learning model can take a handful of numeric inputs and return a risk flag. Every number on this page, including the training set, is invented.
This is not a medical product, it does not give a diagnosis, and it is not affiliated with any clinic, university, or other institution. It must not be used for clinical decisions.
Try a synthetic case
Load a sample or edit the values, then run the demo model. Nothing is sent beyond this app, and nothing is stored.
Result
Run the model to see a synthetic risk flag and probability.
How this demo is built
- 2,500 fully synthetic rows generated by a published script, with a fixed random seed.
- A logistic regression pipeline (standard scaling plus a linear classifier) trained on 75% of that table.
- A FastAPI service and a static front end, packaged in Docker, with automated tests and CI.
- No real clinical records, no private research files, and no institution names.