The architecture
The platform ingests a large volume of market data at high frequency, processing it through a deep neural network to identify small, repeatable signals across a broad universe of liquid equities. No single position carries meaningful weight — diversification is a core structural property of the system.
The strategy is market neutral: the system holds long positions in individual stocks and selected ETFs, while shorting other ETFs. This is used to offset macro-market gamma. The return is driven primarily by the signals themselves. The model is evaluated against live market conditions on an ongoing basis.
Backtested over 25 years of market history and over 12,000 tickers.
