The article analyzes China’s massive e‑commerce market, its rapid mobile‑first growth, distinctive consumer habits, and how platforms like Alibaba differ from Western rivals such as Amazon, offering insights into future global retail trends.
This paper proposes CAIG, a novel method for generating high-CTR advertising images using multimodal large language models, combining reinforcement learning and preference optimization to align generated content with product features.
The document describes the evolution, design principles, key technologies, online inference workflow, evaluation methods, and sample‑generation techniques of a large‑language‑model‑based multi‑agent system that powers a 24/7 e‑commerce merchant assistant, highlighting its benefits, challenges, and future work.
This article details the architectural evolution of the Auto Home e‑commerce platform from its early monolithic stage through micro‑services, master‑data integration, and finally a platform‑centric architecture, describing the technical challenges, design decisions, service orchestration, extension‑point mechanisms, configuration management, tooling, data visualization, and knowledge‑sharing practices that enabled scalable, high‑performance online car sales.
The session presented by senior architects Wang Song and Li Yan detailed the design, technology selection, and implementation of the Car Mall red‑packet system and the Auto Home Double‑11 main‑venue platform, focusing on high‑traffic handling, server‑side rendering, performance gains, and lessons learned from real‑world deployment.
This article explains how Playwright can be integrated with testing frameworks, CI/CD pipelines, monitoring, logging, build tools, API testing, reporting, and distributed testing, and provides a complete Playwright‑pytest Python example that automates user registration, login, product browsing, ordering and post‑order processes for an e‑commerce site.
Poco2 is an open-source Python-based mobile automation testing framework that simplifies UI interaction and validation for Android/iOS apps, demonstrated through a complete e-commerce purchase flow test script covering login, browsing, cart, payment, return, refund, and logout.
This article explains why full‑link performance testing is essential for e‑commerce systems, provides step‑by‑step Python Locust scripts to simulate user actions such as login, product browsing and ordering, shows how to run the tests, and demonstrates multi‑domain testing with weighted tasks.
This article examines how Kuaishou’s e‑commerce platform leverages AI to analyze and improve B‑side user experience, detailing current challenges, the CPS metric model, a self‑built intelligent inspection platform, workflow automation, and future prospects for AI‑driven UX optimization.
The paper introduces ROPE, the first large‑scale cross‑domain product recognition dataset covering detail pages, short videos and live streams, and proposes COPE, a dual‑tower multimodal model that learns unified product embeddings using contrastive and classification losses, achieving superior retrieval and few‑shot classification performance across domains.
Tech Insight presents concise overviews of ten JD retail technology papers accepted at top AI conferences in 2024, covering topics such as open‑vocabulary object detection, multi‑scenario ranking, diversity‑aware re‑ranking, a diversified product search dataset, semi‑supervised query classification, plug‑in CTR models, and methods to mitigate LLM hallucinations.
JD's retail supply chain team presents a data‑driven framework combining inventory selection and allocation algorithms—ML‑Top‑K, Reverse‑Exclude, and an end‑to‑end multi‑task learning model—that improve local order fulfillment, reduce stockout loss and allocation costs, and have been deployed across its RDC/FDC network.
This article examines how large pre‑trained models such as GPT‑4 and BERT are integrated into modern recommendation systems, detailing their advantages, implementation strategies, real‑world e‑commerce case studies, and the technical and privacy challenges that must be addressed for effective deployment.
The article describes JD Retail's AI‑driven multi‑agent platform that models real‑world merchant decision‑making with ReAct‑based LLM agents, detailing the system architecture, agent roles, reasoning loops, workflow examples, training pipelines, monitoring, and future directions for e‑commerce support.
This article presents JD Retail's reverse order fulfillment platform, detailing the business background, product and system support, comprehensive domain‑modeling methodology, challenges, verification, and the practical B‑PaaS engineering architecture that enables high‑concurrency, multi‑role processing and streamlined maintenance of complex legacy systems.
The article explains how JD’s retail user growth team built a data‑driven, AI‑powered “machine” that automates user insight, operation planning, conflict resolution, external touchpoint, and real‑time strategy engines to achieve precise, large‑scale user acquisition and retention.
This article reviews the limitations of traditional DMP and AI‑driven intelligent targeting in e‑commerce, introduces JD.com's Smart DMP framework that combines merchant intent with high‑relevance modeling, and presents experimental results showing over 15% CTR improvement and widespread merchant adoption.
This article systematically explains how to enhance e‑commerce platform availability by implementing both black‑box monitoring to detect functional failures and white‑box monitoring to pinpoint root causes, detailing core order‑process metrics, common issues, mitigation strategies, and illustrative Grafana dashboards.
This document presents a comprehensive solution for merchant settlement in joint‑venture (co‑operated) offline stores, describing business models, settlement subject abstraction, billing engine components, settlement workflow, payment collection, and reconciliation architecture with detailed tables and diagrams.
The article chronicles JD.com's product system evolution—from early SQL Server replication bottlenecks, through Redis‑based read services and full service‑oriented redesign with MySQL clustering, to a decoupled, multi‑layered platform that achieves high availability, scalability, and operational efficiency for billions of SKUs.