An AI trading engine
Trades a fund’s portfolio autonomously: selects trades, manages risk, reports every four hours, and learns from its own results weekly.
A venture fund wanted a systematic trading operation without a trading desk: strategy research, execution, risk, and reporting, all autonomous, with an audit trail an investor could trust.
A 104-module platform: twenty-plus systematic strategies with market regime detection, a backtesting engine, disaster floors and direction-concentration caps, and a two-machine production deployment. On top sits an AI advisor that originates investment theses on a four-hour cadence, attributes every outcome back to the specific knowledge that produced it, and retrains itself weekly.
The platform maintains a public verified track record that is never retro-fitted: projections are stamped before the fact and compared against actuals daily. Informed by its knowledge loop, the advisor’s decision quality improved by an order of magnitude over its blind baseline.
