English:
I'm a Data Scientist and Kaggle Competitions Expert with expertise in Optimization, Game Theory, and Control Algorithms. My work spans Machine Learning, Deep Learning, and Reinforcement Learning, combining mathematical foundations with practical AI applications. Recently I focus on Claude Code Skills, MCP Servers, Edge AI, and Quantitative Finance — building intelligent systems that solve complex decision-making problems in finance, robotics, and the real world.
简体中文:
我是一名数据科学家和 Kaggle 竞赛专家,专长于最优化设计、博弈论和控制算法。我的工作涵盖机器学习、深度学习和强化学习,将数学基础与实际 AI 应用相结合。近期聚焦于 Claude Code 技能包、MCP 服务器、边缘 AI 和量化金融 —— 构建智能系统来解决金融、机器人和现实世界中的复杂决策问题。
|
Competitions Expert · 竞赛专家 |
Convex Optimization · 凸优化 |
Nash Equilibrium · 纳什均衡 |
|
Optimal Control · 最优控制 |
Policy Gradient · 策略梯度 |
Supervised / Unsupervised · 监督/无监督 |
|
Claude Code 金融数据技能包,220+ 接口,用自然语言获取 A 股行情、财务、宏观数据。 Financial data skill for Claude Code — 220+ APIs for A-share market via natural language. |
OpenClaw Windows 一键离线安装包,不依赖海外网络,解压即用。 One-click offline installer for OpenClaw on Windows — no overseas network needed. |
|
面向数据科学的 Claude Code AI Agent 团队,覆盖分析、建模、可视化全流程。 A specialized team of Claude Code AI agents for data analysis, ML & visualization. |
📈 SharpETF基于 Tushare API 的量化投资组合优化工具,以夏普比率最大化为目标求解最优 ETF 组合。 Quantitative portfolio optimization — maximizing Sharpe ratio to find optimal ETF allocations. |
|
面试反作弊朗读检测:韵律特征 + LightGBM 与冻结 wav2vec2 双轨方案,FastAPI 服务,说话人不相交基准。 Audio-only read-speech detection for interview anti-cheating — dual-track prosody+LightGBM / wav2vec2, FastAPI service. |
ML/DL 算法可视化与实战平台 · 60+ 个模型(Transformer / Diffusion / GAN / GCN…)· NumPy 手写 + 框架双实现 · 沉浸式交互动画。 Interactive ML/DL visualization & from-scratch platform — 60+ models with dual NumPy/framework code and live animations. |
|
|
|
|
⭐️ From StanleyChanH

