Small, working instruments that run entirely in your browser — no data leaves the page. Each one turns a piece of my cybersecurity and machine-learning research into something you can actually touch. More are in progress; the first is live below.
Tune a malware classifier’s decision boundary and watch the false-positive trade-off update live — the core contribution of my MSc dissertation.
Open tool →Toggle the nine process-memory feature families on and off and measure how much each one contributes to detection accuracy.
Coming soonA force plot showing which memory features pushed a sample toward “malware” — making the model explain its own verdict.
Coming soonFeed a memory feature vector to the ANN running in-browser and get a malware / benign verdict with a confidence score.
Coming soonHide an AES-encrypted message inside an image with LSB steganography, then extract it back out — all client-side.
Coming soonPerturb a sample’s features and watch it slip past the classifier, illustrating the robustness limits of the model.
Coming soon