Upwork profile — copy & paste

Private draft for setting up your profile. Not linked in the nav. After approval, set SOCIAL.upwork in src/config.ts so Hire buttons point to your live URL.

Profile title

Senior Full-Stack (React/Next.js) | AI Integration, RAG & Streaming Chat

Overview

I'm Mingqi Hou — a senior full-stack engineer with 7+ years shipping production web apps for product teams and startups.

I help clients who need more than a landing page: React/Next.js features that hold up in production, performance fixes (Core Web Vitals, SSR/hydration), and AI that actually works in your product — LLM APIs, streaming SSE chat, RAG with vector search, and agent-style workflows.

━━ What I deliver ━━
• Integrate OpenAI, DeepSeek, Gemini, Qwen, or Ollama into your app (BYOK, encrypted config, session history)
• Build or refactor React / Next.js / TypeScript frontends with Tailwind and solid UX
• Node.js APIs (REST, GraphQL), Postgres, Redis — plus FastAPI when Python fits better
• RAG pipelines: loaders, chunking, embeddings, Milvus/pgvector, evaluation-minded setup
• Code reviews, architecture advice, and clear async written updates (daily or per milestone)

━━ Proof of work ━━
Portfolio site: https://houmq.dev
Open-source AI workspace (demo + code): https://github.com/MingQi39/my-ai-studio
Writing on agents, RAG, and production AI: https://houmq.dev/writing

━━ How I work ━━
• Scope in writing before coding; flag risks early
• Milestones for fixed-price work (1–8 weeks typical)
• Based in China (UTC+8) — good overlap with US/EU mornings for handoffs
• Fluent technical English; prefer async messages over long calls

━━ Good fit ━━
You have an existing codebase, a clear problem, and care about maintainability — not the cheapest "build me a website from scratch in 48 hours" bid.

━━ Not a fit ━━
Pure WordPress/theme-only gigs with no engineering depth, or requests to bypass platform payment.

Message me with your repo (or screenshots), stack, and deadline. I'll reply with questions or a concrete plan.

Categories (pick 1–2)

• Web Development
• AI & Machine Learning

Skills (add in dashboard)

React · Next.js · TypeScript · Node.js · JavaScript · Tailwind CSS · REST API · PostgreSQL · Redis · FastAPI · Python · OpenAI API · LLM · Retrieval Augmented Generation (RAG) · Web Performance

Portfolio 1: My AI Studio — Full-Stack AI Chat Workspace

Built an open-source, production-style AI chat workspace: FastAPI backend (auth, sessions, SSE streaming) and React frontend (Radix UI, Tailwind).

• Bring-your-own-key for DeepSeek, OpenAI, Gemini, Qwen, OpenRouter, Ollama
• Per-user encrypted API config; no secrets in the repo
• 9-language i18n with locale detection and user override
• Session history, system prompts, token-level streaming

Demo GIFs and architecture notes on https://houmq.dev/projects/my-ai-studio — repo: https://github.com/MingQi39/my-ai-studio

Portfolio 2: Production RAG — Ebook Q&A with Milvus

Designed and documented a production-minded RAG flow: document loaders, chunking strategy, embeddings, Milvus vector store, and retrieval tuned for long-form content (ebook Q&A).

Covered evaluation pitfalls, context budgeting, and when vector search beats naive keyword match. Full write-up: https://houmq.dev/writing/production-rag-milvus-ebook

Portfolio 3: Agent & Copilot Patterns for Web Apps

Published deep dives on chatbot vs copilot vs agent loops, plan-and-execute patterns, token/streaming cost control, and shipping AI-assisted coding in real teams.

Helps clients choose the right architecture before over-building. Hub article: https://houmq.dev/writing/chatbot-copilot-agent-loop

Checklist

1. Upload a 1–2 min English video walking through My AI Studio
2. Add houmq.dev and GitHub to profile links
3. Set hourly rate ≥ $30 (or your floor) — avoid racing $15/h website-only bids
4. Enable specialized profiles: Web Development + AI/ML
5. Paste portfolio items above; attach demo GIFs from the site