M37R10N
Research
Factor research, and the MCP process infrastructure that makes model-driven work reproducible.
We build factor signal datasets on the China A-share market, from the topology of equity correlation networks to meaning extracted from non-price data. The research is designed, run, and iterated by LLM agents under senior human direction.
We build Model Context Protocol services that give LLM-native financial and quantitative research a reproducible operating surface: agents can read data, call tools, preserve context, and replay research steps with operational discipline.
We work as a senior professional team with an AI-native research and operations stack.
One market, one team, two research lines. We do not spread across geographies or asset classes.
Every number that leaves this team traces back to versioned source data. If a result cannot be re-derived, it does not ship.
We publish perspective, not specification. Methods, coverage, and performance stay with the licensed product.
A senior professional research team augmented by an AI-driven research and operations stack. This is how the work actually gets done.
Short notes on our two research lines.
Selected public notes on factor research, market-structure indicators, and dataset methodology will be published after internal review.