Constructor

Constructor builds the search and product discovery layer underneath ecommerce - AI-powered, engineered from scratch for retail, not retrofitted from generic enterprise search. The platform processes over a billion queries daily across 150 languages, using machine learning, natural language processing, and transformer architectures to interpret shopper intent and surface the right products. The company's public claim: it has never lost an A/B test against competitive technology. The engineering stack sits at the intersection of search infrastructure and applied ML. Teams work on ranking models, query understanding, and real-time personalization at a scale where latency and relevance directly map to retailer revenue, conversion rate, and profit. This isn't search-as-a-feature bolted onto a catalog - it's purpose-built commerce intelligence. Constructor's culture signals lean toward radical visibility: the company talks about "glass box" insight over black-box opacity, continuous improvement loops, and giving engineers room to operate on the bleeding edge of AI. For anyone whose threat model includes stale ranking algorithms, brittle search pipelines, or ML systems that can't explain themselves, the problems here are concrete and the constraints are real.