Balancing Business Demands and Technical Advancement: Insights from JD’s Data Knowledge Leader Li Wei

Data Value

Li Wei, head of JD’s Data Knowledge Computing Department, explains that JD’s massive e‑commerce data—covering users, merchants, products, and transactions—offers a rich “gold mine” for algorithms that can optimize trading processes, enable precise marketing, and improve user experience.

He describes his team’s work extracting knowledge from diverse data sources (text, images, video) to detect counterfeit or low‑quality products, filter harmful listings, and provide merchants with actionable feedback, thereby empowering the ecosystem.

He also highlights how knowledge‑driven comment filtering can surface high‑value reviews, helping customers make better purchasing decisions.

Thunder Platform: Combating Counterfeits

The “Thunder” platform builds a knowledge base, rule set, and AI models to act as a firewall that blocks counterfeit or illegal items before they reach consumers. By comparing multimodal product information (titles, trademarks, details), the system flags suspicious items for manual review and merchant remediation.

Li acknowledges AI’s limits, noting that human verification remains essential and that a hybrid AI‑human workflow is common across many domains.

Balancing Demand and Innovation

Li stresses the importance of a dynamic equilibrium between fulfilling business requirements and pursuing technical upgrades. Ignoring either side can lead to stagnation; regular technology refreshes aligned with industry standards should be integrated into product development to boost team morale and impact.

He also advocates for deep, regular communication with team members, fostering friendships rather than purely hierarchical relationships to strengthen culture and collaboration.

Weaving Knowledge Networks

Looking ahead, Li sees product knowledge graphs as a strategic focus, aiming to continuously extract, refine, and apply structured e‑commerce knowledge. One application involves linking hot events or celebrity trends to relevant products, creating a closed loop that drives sales through targeted promotions.

Quick Q&A

Li shares personal interests (sports, reading), his view that AI will not surpass humans in the short term, and advice for newcomers to stay abreast of frontier technologies and apply academic learning to real‑world development.