About

I bridge product and engineering — and ship.

A software project lead across design, product, and engineering. I started in UIUX, moved through PM, and now lead delivery — switching fluently between user, ops, and engineering views, with AI embedded in the daily workflow.

Three things I'm built around

  • User empathy — Trained through UIUX work, I surface the real pain behind a request
  • Business translation — From PM years, I translate operational language into engineering specs and cut rework cost
  • Independent delivery — From clarification to diagnosis to cross-team delivery, I close the loop alone

Career path

  • 2022–2024 · UIUX Designer (Allroad Tech) — Learning to read PEOPLE: SaaS commerce, immersive booking, POS systems; user insight & pain-point analysis
  • 2024–2025 · Project Manager (Soohoobook) — Learning to read SYSTEMS: hands-on with SQL, ERD, business logic; business-to-engineering translation
  • 2026 – Now · Software Project Lead (Soohoobook) — Wielding AI & filling tech gaps: Azure CLI self-diagnosis, custom Claude Skills, cross-layer technical judgment

Cross-layer skill map

  • AI tooling design — Claude Code · Cursor · Custom Skills; encoding SOPs into reusable AI workflows so personal know-how becomes team capital
  • Cloud self-diagnosis — Azure CLI · App Insights · Front Door; subscription-level audits and cross-region networking triage
  • DB performance tuning — T-SQL · Query Store · Statistics; distinguishing CPU saturation from lock contention, proposing staged fixes
  • AI product development — Azure OpenAI · RAG · Teams Bot; full product lifecycle from spec to deployment
  • Requirements & design — Figma · Notion · UML · ERD; translating business logic into actionable specs
  • Engineering enablement — Training manuals, SOPs, AI Skills; building a "grep before you assign" / "five-step triage" culture

How I work with AI

An AI coding agent is my daily development interface — but "ask and obey" is not how I use it. Every answer it gives me gets checked against the real data. When they don't match, what I correct isn't just that one answer: I feed back the actual business context the code and the data map to — what a column really means commercially, which number actually reconciles, which pattern has burned us before. Every conversation trains it to reach a valuable answer faster the next time.

What matters more is getting that training out of my chat window and into company assets. I think of it as three layers:

  • The conversation layer — know-how that lives only in one chat, gone when the window closes. A starting point, never the destination
  • The skill layer — validated methods distilled into reusable skills, version-controlled on GitHub, published through branches and PR review, shared by the whole team
  • The tool layer (MCP) — logic that is high-risk, high-frequency, and must return one answer for everyone (like the revenue calculation), locked into diagnostic MCP tools where even the AI can't get it wrong

Across the delivery loop, AI is embedded at every step:

  • Requirement — Deep ops interviews; Claude Code organizes transcripts and assumption lists; MVP scope defined
  • Translation — Figma flows, UI mockups, PRDs; AI drafts UML/ERD, humans validate business detail
  • Dev hand-off — Use established components, pre-validate SQL structure; grep the real codebase first, then produce assignments
  • QA & UAT — Full Use Cases; auto-draft test cases, refine with business knowledge

AI doesn't replace judgment — AI accelerates the execution of judgment.

Why me

  • AI tooling designer — Not just a user of AI. I design AI workflows, build custom Claude Skills, and turn personal know-how into team assets
  • Cross-layer diagnoser — In the gap where no senior Ops exists, I run cross-layer diagnosis and audit reports independently using Azure CLI, Query Store, T-SQL
  • Business-tech translator — Three stages — UIUX → PM → project lead — let me switch fluently between user, ops, and engineering views

Available for

  • Short-term contracts (4–8 weeks) building AI tools
  • Consulting on AI integration into existing products
  • Knowledge base / internal-tool product strategy
  • Earlier design work: design portfolio (archive)

If any of that resonates, get in touch.