Official platform of JD Retail Technology, delivering insightful R&D news and a deep look into the lives and work of technologists.
This article explains the concept of heterogeneous treatment effects, outlines how to select dimensions for HTE analysis, describes a Python‑based MVP tool for automated CATE exploration, and showcases a real‑world experiment case where sub‑group insights turned a non‑significant overall result into actionable findings.
A detailed investigation reveals that MySQL's lock matrix and MVCC‑based transaction isolation caused a concurrent synchronization job to lose metric and dimension data, and proposes a distributed‑lock solution together with a long‑term read‑calc‑write refactor to eliminate the issue.
JIMDB, a high‑performance Redis‑based distributed cache, introduces the “Big‑Hot Key” concept to dynamically identify keys that strain CPU or bandwidth, and implements a multi‑layer active governance framework—including real‑time detection, server‑side caching, circuit‑breaker, and client‑side consistency—to dramatically reduce resource consumption and boost throughput.
This article outlines the motivation, design, architecture, and engineering of an AI-driven experiment analysis agent, detailing its modular workflow, large‑model selection, prompt engineering, front‑end form integration, and future enhancements to improve reliability, transparency, and user interaction.
This article explains how large‑model AIGC technology reshapes e‑commerce by enabling mass‑produced, user‑profile‑driven visual assets, detailing the evolution from early online trade to the 2.0 era, the technical pipeline of multimodal models, and the practical impact on merchants.
JD’s Oxygen Tryon leverages advanced AI, keypoint detection, and real‑time rendering to let shoppers virtually try on clothing, dramatically cutting return rates, boosting conversion, and outlining technical challenges, innovations, and future plans for broader fashion applications.
This article walks through the challenges of building high‑availability systems—covering applications, databases, caches, message queues, containers, GC, and more—using JD’s production experiences to highlight common pitfalls, root‑cause analyses, and practical mitigation strategies for engineers seeking resilient architecture.
The article recounts how a risk‑control algorithm engineer transformed a generic large language model into a specialized e‑commerce fraud detector by iteratively designing prompts, injecting business rules, structuring I/O, and introducing a dual‑hypothesis decision framework to achieve accurate, automated risk analysis.
This article explores the rapid evolution of large‑model AI tools, outlines JD’s open‑source solutions such as JoyAI, JoyAgent, OxyGent and JoyCode, and examines real‑world applications, design principles, policy considerations, and future directions for AI agents and embodied intelligence.
Learn how to quickly set up OxyGent, a flexible AI agent framework, by installing Python, Node.js, and the MCP tools, configuring environment variables, and using just 20 lines of code to build, debug, and deploy multi‑agent applications with features like RAG, tool integration, and distributed execution.
