Works on verifiability of AI outputs and the audit chain behind them: NextModel's receipt system and observability design, from generation to an anchored, inspectable record. Tracks industry protocol work and compliance requirements. Advocates making verification a default capability rather than an afterthought.
专注 AI 输出可验证性与审计链路设计。负责 NextModel 的 receipt 体系与可观测性方案，推动从生成到落地的完整溯源记录。关注行业协议演进与合规要求，撰写的中英双语内容多次被内部审计与外部合作方引用。主张把「可验证」做成默认能力而非事后补丁。
ai-provenance · llm-observability · receipt systems