Quant Research · AI Systems · Backend / Full-Stack
量化研究 · AI 系统 · 后端 / 全栈开发
我喜欢把一个模糊的想法拆成可以运行、可以检查、也可以继续迭代的东西:先弄清楚问题和数据,再把实验做成代码,最后整理成别人能用的工具。
I like turning vague ideas into things that can be run, checked, and improved: understand the question and the data first, turn the experiment into code, then package it into something useful.
目前主要在做三类项目:
- Research systems / 研究系统 — market data, validation, replay, backtests, reports
- AI & backend tools / AI 与后端工具 — agents, APIs, workers, dashboards, automation
- Small products / 小型产品 — study tools, bots, and practical web applications
🎯 正在寻找量化、AI 或后端开发实习。
Currently looking for a Quant, AI, or Backend internship.
Question → Data → Checks → Experiment → Review → Ship
问题 → 数据 → 校验 → 实验 → 复盘 → 上线
我更在意过程是否清楚、结果能否复现,以及失败时能不能知道问题出在哪里,而不只是展示一张漂亮的结果图。
I care about a clear process, reproducible results, and knowing why something failed—not just showing a pretty chart.
| Area | Tools | 方向 |
|---|---|---|
| Core | Python · TypeScript / JavaScript · Go · Java | 后端、脚本、全栈项目 |
| Research | Market data · Backtesting · CatBoost · BC / PPO | 数据研究、策略实验、模型验证 |
| Backend | FastAPI · Workers · REST APIs · Telegram bots | API、后台任务、消息推送 |
| Frontend & Delivery | React · Vite · Docker Compose · GitHub Actions · GitHub Pages | 控制台、部署、CI |
这不是我的真实日常照片,而是把我做项目时常见的一条路径画成了一个小故事:发现问题 → 清理数据 → 质疑结果 → 做成工具 → 留下下一个问题。
This is not a literal diary. It is a visual metaphor for a typical project loop: find a problem → clean the data → challenge the result → build a tool → find the next question.
| Project | What it shows / 项目亮点 | Stack |
|---|---|---|
| poly_strategy | Prediction-market research pipeline: collection, relation discovery, replay, backtest and dry-run checks. / 预测市场研究链路:采集、关系发现、回放、回测和干跑验证。 | Python · Polymarket · Kalshi |
| Atlas20 | Point-in-time rotation research with reproducible backtests, a FastAPI worker and a React/Vite console. / 可复现的轮动回测、FastAPI worker 和 React/Vite 控制台。 | Python · FastAPI · React · Docker |
| Crypto_Research_Agent | An autonomous research workflow with public data, validation, reports and paper-simulated experiments. / 用公开数据做验证、报告和纸面实验的研究 Agent。 | Python · Agents · SQLite |
| crypto-alpha-portfolio | A structured way to test token-unlock, wallet, funding and anti-Sybil hypotheses before putting them into a portfolio. / 系统研究 token unlock、钱包行为、资金费率和反女巫假设。 | Python · Research · Backtesting |
| meme | BSC token lifecycle data → dataset → hybrid model → bot, with explicit paper/live safety boundaries. / 从链上生命周期数据到数据集、混合模型和 bot,并明确区分 paper / live。 | Python · BSC · CatBoost · RL |
| LTT_Strategy | A multi-timeframe signal monitor using Bitget/Yahoo data and Telegram alerts. / 多周期信号监控,连接 Bitget、Yahoo Finance 和 Telegram。 | Python · Telegram · APIs |
- ai-practice-study — Interactive AI revision site with 51 topics and 130 questions. / 51 个考点、130 道题的 AI 复习站。
- polyFIFA2026 — Football prediction research prototype. / 足球预测研究原型。
- ArenaFIFA2026 — Football data research project. / 足球赛事数据研究项目。
- fundingRate — Funding-rate arbitrage research. / 资金费率策略研究。
- second-hand-trading — Java full-stack second-hand marketplace project. / Java 全栈二手交易平台。
- KXO325 — Bilingual logistics study guide with a live demo. / 带在线演示的双语物流复习站。
如果你在招 量化、AI 或后端开发实习,欢迎直接发邮件:
If you are hiring for a Quant, AI, or Backend internship, email me directly:
也欢迎通过 GitHub 联系我:
You can also reach me through GitHub.
Thanks for stopping by. / 谢谢你看到这里。











