Details for the Nerds

Technical architecture reference -- more sections coming soon.

Coming Soon

This page will contain a full technical breakdown of the AI Malware Guardian detection pipeline, including:

• Static PE autoencoder architecture (layer dimensions, training corpus, threshold calibration)
• ETW behavioral feature extraction (12 signals, 5-minute window design, normalization)
• Behavioral autoencoder training methodology and anomaly scoring
• False positive rate analysis and sensitivity tuning
• ONNX runtime inference pipeline
• Alert tier classification logic

Check back as the project progresses toward public release.

Back to home