Calybris PlasmaGuard · Fusion Safety

Disruption risk,
decided safely.

PlasmaGuard learns disruption risk from real FAIR-MAST tokamak shots, quantizes it in a Rust replay engine, and turns it into auditable CONTINUE / MITIGATE / ABORT decisions — uncertainty-aware and fail-closed.

continue mitigate abort
01 — THE PROBLEM

In fusion, an unsafe call is a wrecked machine.

A tokamak disruption can damage a device that costs a fortune and months to rebuild. An ML model can estimate disruption risk, but a floating-point probability is not a safety decision — and a model that's confidently wrong under uncertainty is worse than none. What's missing is a control layer that is deterministic, replayable, and biased toward safety.

02 — HOW IT WORKS

Estimate → quantize → decide.

01 · TAHMİN ET

Conformal ML

A model estimates disruption risk with a conformal upper bound — uncertainty made explicit.

02 · SABİTLE

Rust replay engine

Risk is quantized from float to fixed-point bps, so the same shot always replays into the same chain.

03 · KARAR VER

Fail-closed verdict

If the conformal bound crosses the threshold, it moves toward MITIGATE or ABORT — never relaxes.

03 — WHAT IT SHOWS

Calybris, in a domain outside finance.

Real shots

Built on public FAIR-MAST MAST tokamak data — not synthetic telemetry.

Fail closed

Under uncertainty the system tightens toward MITIGATE, never toward risk.

Replayable chain

Every shot reproduces the exact same auditable decision chain.

A safety-critical decision problem?

Let's talk about deterministic, replayable, fail-closed decisioning.

Start a conversation →