I'm Qingshiyuu, a full-stack developer interested in turning small technical ideas into real, deployed products.
Right now my work is centered around three directions:
- AI developer tools: repo analysis, RAG demos, GitHub MCP experiments, code intelligence
- Production web systems: Spring Boot, React, MySQL, Redis, OAuth, deployment, CORS, cache design
- Cloud and infrastructure: AWS, ECS, Terraform, serverless architecture, CI/CD
I like projects that go past "it works locally"
and survive the messy path to production.|
A GitHub trending intelligence platform that renders README files, generates AI repo briefs, supports auth/favorites, and runs as a real full-stack deployment. Stack: Spring Boot, React, Vite, MySQL, Redis, Railway, Cloudflare Pages |
Python experiments around repository analysis and AI-assisted code understanding. This is part of my longer-term direction: tools that help developers understand unfamiliar codebases faster. Stack: Python, LLM workflows, repo intelligence |
Based on my latest public repositories, my work is moving through a few connected tracks:
| Track | Repositories | What it shows |
|---|---|---|
| Product engineering | github-trending-monitor, QingSH-J |
Full-stack product work, profile polish, deployment |
| AI and agents | repo-agent, simple_rag_demo, github_mcp, github-mcp |
RAG, MCP, GitHub automation, repo intelligence |
| Cloud infrastructure | aws-learn-website, aws-ecs-architecture, aws_exec, aws-resume, serverless_article_with_aws |
AWS, Terraform/HCL, ECS, serverless, cloud learning |
| Backend fundamentals | bloom_filter, tcc, distributed_leak_bucket |
Distributed systems concepts, rate limiting, transactions, Java backend ideas |
| Coursework and research | cs5483, cs6290, CS5481Project |
Python, notebooks, computer science coursework |
flowchart LR
idea["Idea"] --> prototype["Prototype"]
prototype --> product["Usable product"]
product --> backend["Backend contracts"]
backend --> data["MySQL / Redis"]
data --> deploy["Cloud deployment"]
deploy --> lessons["Write the lessons down"]
I care about the parts people usually meet after the demo works:
- auth flows that actually redirect correctly
- cache design that protects API cost
- CORS and environment variables that behave differently in production
- database initialization and migration strategy
- frontend routing that survives refreshes and custom domains
- Shipping Repo Radar as a polished developer tool
- Improving AI summaries with cache, quota, and rate limiting
- Learning cloud-native deployment through AWS, Railway, and Cloudflare
- Building small backend demos for distributed systems concepts
