Products
Learned surrogates.
Built for your physics.
We build validated surrogates of your simulator and deliver them as managed endpoints — each with a published validity report: the classical baseline it had to beat, calibrated uncertainty, and the regions where it abstains. Fluxus is what that looks like in production for superconducting electronics.
01 · Custom surrogates — our core offering
You bring the domain. We build the model.
For simulation problems no off-the-shelf tool touches: we train a surrogate on your simulator, calibrate it to your process rules, validate it against held-out ground truth, and hand you millisecond-latency predictions with confidence bounds on every output. Fluxus, below, is what that looks like in production.
Superconducting design studio
Fluxus
Design the whole superconducting stack
Fluxus pairs an analytic cQED oracle with neural surrogates to design transmon-resonator qubits and RSFQ logic in one studio — Hamiltonian design, chip-yield Monte Carlo, inverse design, floorplan co-design, and a JoSIM-free RSFQ margin engine. Every prediction carries a calibrated interval and a source label.
Run as an API or CLI, Fluxus batch-verifies a cell library against your PDK at 200–500× SPICE, with low-confidence cells auto-flagged for a full SPICE check. Validated on MIT-LL SFQ5ee+; trained per process node on your reference simulator.
RSFQ verification · at chip scale
chip-scale
held-out chains
held-out chains, RMSE
held-out cells, grouped split
every output
low-confidence
Process nodes
02 · Evidence — see it before you buy it
Don't take our word for it.
Two suites are live on our demo portal — superconducting logic screening and lithium-ion degradation — built with the same pipeline we would point at your simulator. Every prediction carries a risk percentile from twin-model disagreement, and each model abstains in the regions its reference physics cannot label. The validity reports, including the classical baselines our own models had to beat, ship with the models.
RSFQ screening — Fluxus
SuperconductingMargin, timing, yield, and design-rule screening for RSFQ logic cells — layout-native, validated on MIT-LL SFQ5ee+.
100% DRV detection · 99.1% chain functional match, held-out
~10³× vs. JoSIM margin runs
Battery-life screening
Lithium-ion500-cycle degradation from 11 cell-design and duty knobs — an emulator of PyBaMM aging simulations for design-for-longevity sweeps.
0.49% retention-curve MAE · 91.0% conformal coverage @ 90% target
~5×10⁵× vs. the aging simulation
Your physics. Our surrogate.
Bring us a simulation bottleneck in any physics-constrained engineering domain — superconducting electronics, biotech, robotics, or something we haven't seen yet. We want to understand it.