Machine learning and explainable AIIndependent founder project
OncoView AI
An explainable medical-imaging triage demo: an ensemble model + Grad-CAM shows the evidence behind each decision.
Problem
A medical-imaging model answer is hard to trust when you cannot see what it looked at.
Founder’s role
Dmitri built an ensemble model and a Grad-CAM explanation that shows the evidence for each decision.
Result
Each prediction shows what the model relied on in the image (a public demo, not a diagnosis).
Problem
Imaging models are often shown as a probability with no visual evidence, no operating-point choice and no public way to test behaviour - a poor fit for a screening-style discussion.
Solution
A demo where the user picks a case or uploads an image, chooses balanced or high-recall mode, and inspects the prediction, confidence and heatmap overlays on one screen.
Result
A live explainable demo: ensemble inference, two modes and visual evidence. It is a research demo, not a diagnosis.
