AI in practice
AI-assisted development, agentic workflows, MCP, RAG, and LLM integrations. Exploring responsive AI experiences with streaming UI and the Vercel AI SDK.
I'm Lewis Liu, a senior full-stack engineer at HPE. I build AI-powered experiences and use agentic workflows to connect ideas with dependable, production-ready software.
Zhonghui Liu · Building across the full product lifecycle

Agentic workflows · MCP · RAG · LLM integration · Vercel AI SDK
Python · Django · FastAPI · REST · GraphQL · SQL
AWS · Docker · Kubernetes · CI/CD
React · Next.js · TypeScript · Design systems · Accessibility
Senior Full Stack Engineer · 2020–Present

My current focus connects AI-powered experiences with the backend services, cloud infrastructure, and architectural decisions that support them.
Agentic workflows, MCP, RAG, and LLM integration, supported by AI-assisted development.
Python and FastAPI services, API integration, and automated testing.
AWS, containers, Kubernetes, and CI/CD for dependable application delivery.
Connecting interfaces, services, and data with maintainability and clear system boundaries in mind.
My current focus is AI-augmented development and AI-powered applications: agentic workflows, MCP, RAG, LLM integrations, and interfaces that make these capabilities useful to people.
I've been building for the web since the early 2000s. That foundation in frontend architecture, full-stack systems, testing, and cloud delivery shapes how I bring AI into practical products—with quality and maintainability built in.
AI-assisted development, agentic workflows, MCP, RAG, and LLM integrations. Exploring responsive AI experiences with streaming UI and the Vercel AI SDK.
React, Next.js, TypeScript, accessible interfaces, design systems, and clear user experiences. Backed by automated tests and thoughtful architecture.
Node.js, Python, Java, REST and GraphQL APIs, SQL and NoSQL databases. Cloud delivery with AWS, Docker, Kubernetes, and CI/CD.
Selected projects across public services, university operations, commerce, and productivity—spanning interfaces, APIs, data, and delivery.

Full-stack contributions to OCAT: core interfaces, REST and GraphQL APIs, data models, automated tests, and accessibility improvements.

Full-stack development for the Admin-IT TS Work Order system, spanning UI/UX, GraphQL services, automated tests, and database refactoring.

Core frontend development for CANS, with leadership on data visualization, printing functionality, and automated acceptance and regression testing.
Explore project repository ↗
Grocery shopping interfaces for Haigy, restaurant ordering and management for L&M, and Vlink, a visual bookmarking prototype.
From frontend and full-stack foundations to AI-assisted delivery and intelligent product experiences.
Current focus: AI-augmented development, MCP and agentic workflows, RAG and LLM integration, streaming UI, cloud applications, and automated testing.
Admin-IT TS Work Order system.
CANS and OCAT public-sector applications. Cambria Solutions was acquired by EY in 2022, after my tenure.
Commerce platforms, restaurant management, prototypes, and local business websites.
Web interfaces, interactions, and data visualizations for Ctrip, ToursForFun, and HHtravel.
Education resource platforms and Shandong University web services.
Research Scholar · Mathematical analysis of cognitive labor organization
Bachelor's degree · Computer Science
Languages: Chinese · English
AI-powered applications, agentic workflows, or the engineering foundations behind them.
Let’s connect.