「Turning market patterns into code, and ideas into products.」
Futures Trader × Full Stack Engineer × AI Agent Builder
I'm a full-stack developer with years of hands-on experience in futures trading and deep expertise in Chan Theory (缠论) technical analysis. I build quantitative trading systems from scratch — from K-line merging and fractal recognition to real-time alerting and strategy backtesting. Recently I've been diving into AI Agents and LLM applications, exploring how language models can automate trading analysis and cross-device workflows.
- 🎓 HPU — Henan Polytechnic University
- 📈 Futures Trader — Chan Theory practitioner, self-built quant system
- 💻 Full Stack — Vue / React / TypeScript / Node.js / FastAPI / Python
- 🤖 AI Agent — LLM tool calling, RAG, multi-turn conversation systems
- 🔗 Cross-Device — WebSocket-based PC ↔ mobile real-time synchronization
- 🌱 Project-driven learner — I ship, therefore I learn
| 🖥️ Full-Stack Development | Building complete web applications from pixel-perfect UIs to scalable APIs. Vue / React ecosystem on the frontend, Node.js & FastAPI on the backend. |
| 📊 Quantitative Trading | Designing and implementing Chan Theory-based trading algorithms — K-line consolidation, fractal identification, stroke (笔) generation, multi-timeframe analysis, and full backtesting engines. |
| 🤖 AI Agent Engineering | Developing LLM-powered agents with Function Calling, RAG knowledge bases, and structured tool use. Connecting AI to real-time market data and trading workflows. |
| 🔄 Real-Time Systems | WebSocket bidirectional communication, live market data streaming, multi-client synchronization, and PC-to-mobile cross-device connectivity. |
Frontend
Backend & Data
AI / ML / Quant
Tools & Platforms
My flagship project — a complete Chan Theory futures quantitative system built from the ground up:
| Component | Details |
|---|---|
| Chan Algorithm | K-line inclusion merging → fractal identification → stroke (笔) generation. Verified bar-by-bar against the czsc library with 100% consistency. |
| Market Data | TqSdk persistent connection, 17 futures contracts, all timeframes from 30s to daily. |
| Real-Time Alerts | FastAPI + WebSocket bidirectional sync across 4 pages — PC overview / detail / backtest + mobile client. |
| Backtesting | Standalone backtest engine. Fills at fractal confirmation bar close (no look-ahead bias). Equity curve + trade markers + per-trade breakdown. |
| Risk Management | Simulated 1000k account, 1% risk per trade position sizing, drag-and-drop stop-loss calculator. |
| Charting | lightweight-charts professional financial charts with multi-timeframe switching and Chan stroke overlay. |
| Area | What I'm Building |
|---|---|
| 🤖 AI Agent | LLM-powered multi-turn agents with Function Calling, RAG vector knowledge bases, and market data tool integration |
| ⚡ FastAPI Services | High-performance async Python backends with auto-generated OpenAPI docs, serving AI model inference |
| 📈 Chan Theory Deep Dive | Segment (线段) modules, pivot (中枢) identification, sub-level interval nesting, structural stop-loss |
| 🔗 Cross-Device Sync | PC FastAPI server + mobile PWA over WebSocket, sensor data streaming, remote device control |
| 🧪 Strategy Evolution | Moving from single fractal signals to multi-factor, multi-timeframe resonance strategies |
- 🥇 2022 National Mathematical Modeling Competition — China (MATLAB)
- 📊 MathorCup University Mathematical Modeling Challenge — Big Data track
- 🧠 BP Neural Network Prediction — Self-built forecasting model
- 📦 64+ GitHub Repositories — Frontend, Backend, Quant, AI, Blockchain, Mobile
- 🎯 17 Futures Contracts — Real-time monitoring and alerting
- 📱 PC + Mobile Dual-Platform — WebSocket synchronized trading dashboard
JavaScript ████████░░░░░░░░░░░░░░ 31.7% (19 repos)
HTML ████░░░░░░░░░░░░░░░░░░ 18.3% (11 repos)
Vue ███░░░░░░░░░░░░░░░░░░░ 11.7% (7 repos)
Python ██░░░░░░░░░░░░░░░░░░░░ 6.7% (4 repos)
MATLAB █░░░░░░░░░░░░░░░░░░░░░ 5.0% (3 repos)
CSS █░░░░░░░░░░░░░░░░░░░░░ 5.0% (3 repos)
Others █████░░░░░░░░░░░░░░░░░ 21.7% (13 repos)
「Ship fast, iterate faster. The best way to learn is to build.」
- Project-driven — I learn by shipping real products, not by watching tutorials
- Full-stack mindset — From database schema to pixel-perfect UI, I own the whole stack
- Trading discipline — Years of futures trading taught me risk management, patience, and data-driven decision making — skills I bring to every engineering project
- AI-native thinking — I design systems with LLM integration in mind from day one, not as an afterthought
- Cross-device obsession — I believe your phone and computer should work as one, not two isolated devices
- 📧 Email: A18581897806@outlook.com
- 💬 WeChat: 18581897806
「Trading is the realization of cognition. Code is the extension of ideas.」
Thanks for stopping by — feel free to reach out! ⭐