Xueqi LiuFull-stack Product EngineerOpen to full-stack / product engineering rolesContact

About

I build software from the decision boundary outward.

The thread across my work is turning ambiguous operations into explicit, testable systems.

01

What I do now

I work as a full-stack product engineer across product definition, system architecture, implementation, and verification. I am most useful where a workflow contains consequential decisions: an approval that cannot be bypassed, evidence that may become stale, a retry that could duplicate a side effect, or private information that must not travel farther than its purpose. My portfolio therefore emphasizes observable behavior and honest boundaries, not just polished screens. I am currently looking for a product-engineering or systems-oriented role where I can own a difficult problem from framing through delivery and make the reasoning behind the solution inspectable.

02

How the capability formed

Before turning these ideas into a public engineering portfolio, I worked as a technology lead within a privately held international-trade business. That operating context made software consequences concrete: research competes for time, working capital is limited, supplier and logistics information changes, and an apparently small automation can create a costly commitment. It taught me to model uncertainty before optimizing a workflow and to treat exceptions as first-class product states. I do not publish the company’s identity, customer details, or private operating data. The useful part of the story is the perspective it produced: technology decisions are business decisions, and a system is credible only when its authority, evidence, and recovery paths are explicit.

03

What I am exploring

My current work connects three themes. TradeFoundry explores how quantitative disciplines—evidence lineage, hypothesis ranking, constrained allocation, and feedback—can improve global-trade decisions without pretending to predict outcomes or authorize transactions. CargoMesh and the n8n Reliability Lab examine durable execution, verification states, idempotency, and recovery when independent systems disagree. Women AI Network explores privacy as an API and persistence contract rather than a preference hidden in the interface. Together these projects express the kind of engineer I want to be: commercially aware, precise about claims, comfortable with ambiguity, and willing to carry responsibility through testing, documentation, and release readiness. I publish the decision patterns behind the work so a hiring team can evaluate judgment, not merely count frameworks.

From narrative to evidence

Inspect the projects and thinking directly.

ProjectTradeFoundryProjectWomen AI NetworkEssayWhat quantitative thinking contributes to global trade

Ownership

How I own the work

Ownership means keeping the problem, constraints, decisions, evidence, and release boundary coherent—even when implementation uses AI assistance and mature open-source components.

  1. 01

    Define

    I frame the user, decision, failure cost, privacy boundary, and acceptance criteria.

    AI may help challenge assumptions; it does not choose the product claim or authority boundary for me.
  2. 02

    Design

    I choose system boundaries, data ownership, failure behavior, and the evidence required to support public claims.

    I reuse proven open-source infrastructure when its license, maintenance burden, and behavior fit the constraint.
  3. 03

    Implement / review

    I integrate the system and review high-risk paths, contracts, migrations, and irreversible effects.

    AI tools can implement routine, bounded work; generated code is not accepted until it fits the architecture and passes review.
  4. 04

    Verify

    I define negative cases, run checks, inspect failure modes, and separate execution from proven outcome.

    Automated tests and independent review supply evidence; they do not turn a local fixture into a production claim.
  5. 05

    Deliver

    I decide what is ready to publish, document limitations, and preserve a recoverable release path.

    Deployment tooling performs the release; I remain accountable for the public boundary and acceptance decision.

Now

Current focus

A small, dated list of what I am actively prioritizing—not a private roadmap.

Product-engineering role

Exploring full-stack and systems roles where product judgment, reliable delivery, and cross-functional ownership matter.

Updated 2026-09-01
TradeFoundry decision system

Refining the private MVP and its public, evidence-bounded commercial narrative for global-trade decisions.

Updated 2026-09-01
Decision-system essays

Publishing the reasoning patterns behind verification, quantitative trade research, and privacy-by-contract.

Updated 2026-09-01