Zalando Search: Architecture Behind Product Discovery - Ivan Potapov & Waleed Tariq - MICES 2026

Inside Zalando Search: Architecture Behind Product Discovery at Scale A look inside the Search & Browse architecture powering product discovery for millions of customers across Zalando's European markets. We'll walk through the system end-to-end: Elasticsearch clusters with lexical and vector retrieval, the NER-powered query builder that interprets user intent, the Catalog API orchestrating parallel requests, and the ranking stack — where the Algorithm Gateway applies ML relevance models and re-ranking to turn raw search results into a curated shopping experience. We'll also touch on how promotions blending and multi-market isolation fit into the picture. With a product perspective on what these architecture and ranking choices mean for search quality, partner visibility, and the customer experience. Follow up talk on blog post: engineering.zalando.com Ivan is a Research Engineer at Zalando on the Search & Browse team, powering catalog search and product discovery for millions of customers across Europe. He has a deep interest in Code Agents architecture, is an active open-source contributor, and writes at blog.ivan.digital. Waleed is a Search Engineer at Zalando specialising in retrieval, relevance and query understanding at scale. Previously led backend teams shipping technical products and ML pipelines for international e-commerce platforms. MICES - MIx-Camp E-commerce Search https://mices.co