threshold-calibration

Code and experimental artifacts for "Synchronous Online Verification Gating in Semantic Caches" — an empirical study of real-time verifier gating for semantic cache hits, evaluated against static/adaptive-threshold baselines on ~210k real requests across three datasets.

Calibrated decision layer and risk gate for AI coding agents. Routes yes/no, routing and scoring judgments to the Jev System One model at ~$0.0000123 per decision, then enforces the result deterministically through Claude Code PreToolUse hooks. Fitted thresholds, not guessed.

Empirical label-blind anomaly detection on Numenta Anomaly Benchmark streams with chronological evaluation, alert-budget calibration, event metrics, and reproducible research artifacts.