Catalogue / zalando-2025-inventory-optimisation

Zalando splits dynamic inventory optimisation into weekly Databricks and SageMaker demand forecasting and dual offline-online replenishment pipelines

Zalando, 2025. ZEOS dynamic inventory optimisation system.

Organisation
Zalando
Evidence
Reported from production
Source
Building a dynamic inventory optimisation system: A deep dive

Zalando Science Blog, published 2025-06-30

Records

Decision#d1

Split feature engineering into distributed PySpark pre-processing on Databricks and vertical Python transformations on SageMaker

Decision#d2

Use LightGBM with Nixtla MLForecast to combine demand forecast training and inference in one SageMaker job

Decision#d3

Use dual offline and online SageMaker Feature Store tiers and write interactive updates back to the offline store

Decision#d4

Serve interactive partner inventory-setting updates asynchronously through SQS, Lambda and event-stream notifications

Decision#d5

Isolate every pipeline run on dedicated transient Databricks job clusters and SageMaker jobs via zFlow

Sign in to read what was chosen, what was rejected and what it cost, each with its place in the source. An account is free.