A curated set of recent talks and forum appearances. Each card links directly to the original external coverage.
An invited talk at the Shanghai Post and Telecommunications Economics seminar on how agents move from "usable" to "deployable" through L1-L5 safety evaluation, domain simulators, and efficiency-oriented standards.
An invited talk on LLM and agent evaluation, focusing on governance frameworks, risk control, and scenario-driven deployment in the Shanghai AI Pilot Zone.