I'm a Full-Stack & AI Agent Developer with 8+ years of experience, building at the intersection of AI agents and quantitative trading.
- 🦊 Currently building Inalpha — an open-source professional quant agent framework. Agents pick the factors working now to time entries, write full strategies, and evolve them in a sandbox. Every order goes through machine approval — the LLM is never on the order path.
- 🛠️ Also shipped openfinclaw-cli — a one-stop quant-trading AI agent that runs inside Claude Code, Cursor, and 20+ AI agents via MCP.
- 🌱 Exploring the AI Agent Skills ecosystem, protocol standards (MCP / A2A), and Harness Engineering — building AI development harnesses with headless QA, deploy pipelines, and autonomous dev loops.
- 💬 Ask me about AI Agents, factor investing, Next.js, TypeScript, and Python
- ⚡ Fun fact: I believe the right division of labor is — agents do the research, machines guard the orders.
A quant framework where the agent is the researcher, not the trader:
- Factor timing — agents track which factors are working right now and time entries accordingly, backed by a 70-factor library with lineage audits.
- LLM-written strategies — the agent authors complete strategies from a conversation, then backtests them against fresh, multi-market data.
- Sandbox evolution — strategies are promoted, paper-traded, and evolved in isolation before anything matters.
- Machine approval — a deterministic risk gate sits on the order path. The LLM proposes; it never executes.
- Multi-market, audit-grade — crypto, US/CN/HK equities, global indices, FRED macro. Every number traces to a data source with freshness checks.
- Factor timing and factor-library engineering (lineage audits, IC benchmarks, factor discovery)
- LLM-written strategies with audit-grade boundaries
- Strategy evolution in sandboxes — promote, paper-trade, reflect
- Machine approval on the order path (risk gates, permissions, plan-exec)
- AI Agent Skills ecosystem, MCP / A2A, and Harness Engineering
Always happy to talk about AI agents, quantitative trading, and what happens when you put the two together.





