Future of Stats

For researchers, modellers, and statistics experts

Leverage the power of ML to cut down statistical regression and modeling times to seconds.

Have you ever wondered about the statistical uncertainties on your ML model parameters? Do you feel there is a disconnect between ML and statistics libraries? We feel that, too: The arbitrary split slows YOUR analysis workflows and hides valuable information about statistical uncertainties in YOUR data.

We are building a unified statistics engine, where mathematical functions, gradients, minimization, and advanced statistical analysis methods utilize optimal hardware acceleration from the moment you write your models.

Maximum-likelihood fit

converged
β₀1.842 ± 0.011
β₁-0.357 ± 0.008
σ0.204 ± 0.003

Gradient norm

2.7e-9

Backend

GPU

Early access application

Tell us your biggest modeling bottleneck

Which stack are you using most right now?