Whenever you're ready, walk me through how you'd approach this.
Describe your strategy for ensuring data privacy and security when an AI agent interacts with proprietary customer design files.
Discuss the end-to-end architecture, including identity, authorization, retrieval, model inference, tool execution, tenant isolation, auditability, and failure handling. Explain how you would evaluate the design, monitor it in production, and prevent training-serving skew or accidental data exposure.