Fits flexible maximum likelihood regression models supporting censored, interval, and hybrid continuous/dichotomous data. Provides explicit analytic and numerical gradient computation, random intercept models via Gauss-Hermite quadrature, and multiple distribution families.