Manufacturing-export operator building auditable AI workflows.

I work where export operations meet factory reality: customer discovery, RFQs, pricing preparation, trade documents, and the checks that keep small errors from becoming expensive ones.

My focus is not “AI that runs a business by itself.” It is turning repeated, high-consequence work into systems with clear data, deterministic rules, review gates, and accountable human decisions.

• Trade Pipeline — one order record, connected quotation, PI, commercial invoice, and packing list outputs for manufacturing exporters.
• Tianji Skills — open, de-identified workflow components for B2B lead research, RFQ intake, and trade-document verification.

• Evidence before fluency. Missing data is surfaced, not invented.
• Automation with gates. AI can parse and assist; critical identity, quantity, weight, price, and terms stay reviewable.
• Public methods, private operations. Open-source examples use fictional or fully de-identified data—never customer, supplier, pricing, cost, or banking records.

I am interested in conversations with manufacturing operators, trade-tech builders, and teams exploring practical AI workflows for industrial export.

Yang Tian is my Chinese-facing personal identity; Ethan Yang is used in international technical and business contexts. “Dangooy” is the GitHub account behind both.

Building systems that make operational knowledge more transferable, verifiable, and useful—without pretending uncertainty has disappeared.