Zalando Hosts Workshop on Optimization and ML

Per Zalando's engineering blog, Zalando hosted the 106th meeting of the GOR working group "Praxis der Mathematischen Optimierung" on October 6-7, 2022 at Zalando Headquarters in Berlin (GOR invitation PDF; Zalando engineering blog). The event ran in hybrid mode with in-person attendance of about sixty participants and virtual streaming for remote attendees (Zalando engineering blog). Presentations were given by representatives from industry and academia across Germany and covered applications including forecasting, network design, pricing, logistics, scheduling, and vehicle routing (Zalando engineering blog).

Per the Zalando report, speakers addressed problems at the intersection of mathematical optimization and machine learning, and Zalando applied scientists presented two concrete use-cases: the pricing team discussed challenges in large-scale article discounting, and the logistics team presented on stock distribution and related operational challenges (Zalando engineering blog). The GOR invitation frames the meeting as targeted at practical optimization topics and states the working language as English to include non-German speakers (GOR invitation PDF).

Editorial analysis - technical context:

Companies and research groups convening on both optimization and ML typically surface trade-offs between model-driven forecasting and combinatorial decision layers, such as coupling learned demand estimates with integer or network-flow optimizers. Practitioners will commonly discuss topics like robust demand forecasting, scalable approximate solvers for routing, and heuristics that integrate learned components with exact optimization.

Editorial analysis: The workshop sits at a practical intersection for e-commerce operations where both predictive models and optimization algorithms are required to scale decisions across catalogs and supply networks. Public reporting notes Zalando serves a large customer base and handles substantial order volumes, which creates problem instances that benefit from combined ML and optimization approaches (Zalando engineering blog). For data scientists and ML engineers, these settings emphasize production constraints: latency in decision loops, data freshness for forecasts, and solver scalability for combinatorial layers.

Editorial analysis: Observers should track follow-up materials such as slides, code releases, or workshop proceedings from GOR that detail algorithms, benchmark results, or open datasets. Indicators of practical transfer include published case studies on discounting pipelines, open-sourced solver wrappers or approximation routines, and reproducible comparisons of end-to-end metrics (for example revenue uplift or fulfillment lead-time improvements) reported by hosts or speakers.