MonicaWu

I design it and I build it. One person owns the interface, the component system, and the code behind them — from the first prototype to the production deploy.

I'm a design-trained engineer with 10 years shipping production UI end to end. I build design systems other teams build on, and in 2026 I shipped two production AI systems, both with per-stage evals. I work in TypeScript, React, Next.js, Tailwind CSS, Angular, and Python. To me the interface and the data model are one design problem: on the Workbench, the reviewer's screen decided what the data model had to be.

Digital Twin — an agentic RAG agent with structured tool-calling and real-time intent routing (Python, OpenAI, ChromaDB), live on Hugging Face Spaces since May 2026. Reconciliation Workbench — human-in-the-loop AI exception review for multi-jurisdiction tax compliance: 100% recall, zero false positives against seeded ground truth, with a full audit trail.

Software that moves $200M a month doesn't get to break. That's where I learned to ship. Before AI: an enterprise payments platform (Accenture), the design system behind healthcare apps used by physicians and pharmacies nationwide (Inovalon), accessible components on ~50,000 Microsoft partner websites, and clinical research apps across 9 client sites (IQVIA). I took my components through 6 major Angular versions with zero regressions. That's the bar.

I started in design, then Code Fellows, then a decade of production TypeScript, React, Angular, and Node.js. I prototype first, build the real thing, and put it in front of someone. At Inovalon, the component specs I wrote became the frontend standard for every consuming team. In 2026 I cut a human-in-the-loop review product to the one screen a reviewer needs and shipped it in 2 days. I treat AI uncertainty as a UX problem: review surfaces, guardrails, intent routing. I build with Claude and Cursor daily: ~40% faster delivery, every change review-ready.

The AI work started at IQVIA. Our clinical apps had to run locally on client hardware that varied site to site, so I tested whether LLM inference could clear that bar: Llama 3.1 8B on Ollama, constrained hardware, worst case. Self-initiated, no mandate. It set the AI roadmap for regulated clinical deployments.

The conversations I want: senior design engineer, AI frontend, and product engineering roles where the interface and the code behind it are one job — including LLM products shipped to real users. Remote (US), Pacific time.