One person. Four AI departments.
Zero manual data processing.
An AI-native quantitative research firm powered by autonomous multi-agent systems.
Layer 1 — Beta Capture
Risk parity all-weather portfolio across 10 assets: equity index futures, bond futures, gold, and commodity ETFs. Systematic rebalancing with daily monitoring and Telegram-based reporting.
Layer 2 — Alpha Enhancement
23-factor stock selection combined with AI-driven fundamental research on A-shares. Individual stock positions progressively replace passive index exposure to enhance alpha.
Data Infrastructure
28 MCP tools. 5,400+ A-shares, 2,400+ ETFs, futures, full financial statements. Three-tier caching: LRU memory → Tencent Cloud COS (HK) → upstream APIs. Point-in-Time processing.
Factor Research
23 alpha factors with IC/ICIR evaluation. Autonomous factor discovery loop inspired by Microsoft RD-Agent: LLM hypothesizes → implements → backtests → evaluates → iterates.
Agent Orchestration
Four autonomous OpenClaw agents coordinating via MCP protocol. Structured identity, persistent memory, scheduled cron tasks. Extended with external MCP servers for macro research.
Kaihua (William) Huang
6 years in sell-side quantitative finance across derivative valuation, model risk, and curve construction. Now building an AI-native quant firm where autonomous agents do what used to require a team of 20.
Not AI-assisted. AI-native.
Follow on LinkedIn