AI applications and RAG
Embedding, retrieval pipelines, model calling, and evaluation — from experiment to a repeatable service.
ZoeImport / AI Application & Backend Engineer
Zihao Zheng.
About two years of engineering practice across AI applications, backend systems, retrieval infrastructure, and Agent Engineering — built to be inspected, reused, and questioned.
About / 关于我
Small systems, visible behavior, honest claims.
I am Zihao Zheng (ZoeImport), an AI application and backend engineer with about two years of engineering practice focused on AI applications, backend systems, retrieval infrastructure, and Agent Engineering. I studied Computer Science and Technology at Xi’an University of Architecture and Technology (B.S., 2021–2025).
FOCUS
AI applications · Backend systems · Retrieval infrastructure · Agent Engineering
EDUCATION
B.S. Computer Science and Technology, Xi’an University of Architecture and Technology, 2021–2025
Core strengths / 擅长方向
Where I work best.
Backend and data infrastructure
Go services with Gin and GORM, RESTful APIs, and MySQL, Redis, Elasticsearch, and RabbitMQ behind explicit contracts.
Agent engineering
Agent loops, MCP, skills, plugins, tool calling, and context/memory management with observability built in.
Delivery and observability
Docker, Kubernetes, GitHub Actions, and GitLab CI/CD pipelines that make behavior inspectable end to end.
Selected stack / 核心技术栈
Tools I reach for.
- Languages
Go, TypeScript, Rust, SQL
- Frameworks
Gin, GORM, React, Vue
- Data & infra
MySQL, Redis, Elasticsearch, RabbitMQ
- AI / ML
PyTorch, Transformers, Embedding, RAG, model calling & evaluation
- Agent systems
Agent loops, MCP, skills, plugins, tool calling, context/memory, observability
- Delivery
Docker, Kubernetes, GitHub Actions, GitLab CI/CD
- Platform
ArchLinux, systemd
Selected projects / 代表作品
Public work with clear boundaries.
Each case study states what the project solves, its stack, its honest limits, and where the source or case study lives.
SearchX
Provider-aware WebSearch and policy-guarded WebFetch, built as two independently deployable Go services.
Token Atlas
A local-first, privacy-bounded observability dashboard for understanding AI tool activity across multiple coding clients.
Current interests / 当前兴趣
What keeps pulling me in.
- Agent runtimes & harnesses
- Observable event-driven systems
- Local-first AI tooling
- Learning through executable slices