SYSTEM Cited by 1 source

Nixtla MLForecast

Nixtla MLForecast (mlforecast) is an open-source time-series forecasting library from Nixtla that wraps traditional gradient-boosting models (typically LightGBM) with time-series-specific feature engineering and forecast-generation machinery:

• Target lags and lag transformations.
• Exogenous-feature lags and transformations.
• Temporal features (seasonality, calendar effects).
• Conformal inference for probabilistic forecasts with calibrated prediction intervals.
• Numba-accelerated feature generation under the hood for speed.

Stub page — minimum viable framing.

Seen in

• sources/2025-06-29-zalando-building-a-dynamic-inventory-optimisation-system-a-deep-dive — canonical first wiki disclosure as the forecasting library powering Zalando ZEOS's 5M-SKU weekly demand forecast. Verbatim:

"Forecasting specific features such as target lags / transformations, exogenous features lags / transformations, and other temporal features is handled later on by Nixtla's MLForecast. This allows us to leverage optimised transformations from Nixtla (with Numba under the hood)."

And Zalando's architectural framing:

"We leverage Nixtla's MLForecast with conformal inference, and LightGBM for probabilistic forecasts. Beyond the speed and scalability of LGBM, we want to emphasise the benefits of using a library like Nixtla, which can automate many time series features and processes required just before training."

Related

• systems/lightgbm — underlying model.
• systems/numba — JIT accelerator used under the hood.
• systems/sagemaker-training-job — typical Zalando compute host.
• systems/zeos-demand-forecaster — canonical Zalando consumer.
• concepts/probabilistic-demand-forecast
• companies/zalando