Operating System for Structured Credit
Ingest collateral, model waterfalls, and solve capital structures — in one composable engine.
Try these examples
Compare the prospectus default projections to the actual historical default performance of similar auto loan pools in the Edgar Auto Loans database, and highlight any material deviations.
Model the cashflow waterfall from this prospectus and calculate the Class B yield at a price of 104 under a 25% CPR assumption.
Train a logit model to predict auto loan defaults on the Edgar Auto Loans dataset using: vintage = 2021, term ≥ 60 months, FICO < 650. Return AUC, coefficients, and a calibration check.
Apply my default model to this loan tape and report forecasted cumulative defaults and losses—overall and segmented by vintage, term, and FICO band.
Define a structure with a $100mm Class A note paying 5.5% fixed and a residual equity tranche, backed by $115mm of new-issue collateral (17% WAC, 60-month WAM). Excess interest should turbo the Class A until a 20% OC target is reached. Run cashflows and price the equity tranche to a 15% yield assuming 8% collateral losses.
Build your own default models from real loan data
No more relying on third-party models you can't inspect. Train on actual loan-level performance data, see every coefficient, and download the model.
