The Global AI Alpha Fund combines a proprietary six-layer AI engine with experienced fundamental judgment to exploit the gap most funds ignore. We score over 5,000 global securities daily. Every high-conviction position is validated by a human analyst before it enters the book. The result is a quantamental strategy built for enduring, risk-adjusted performance.
Quantitative models process breadth no human team can match. But they are backward-looking by design, blind to structural breaks, regulatory shifts, and the context behind the numbers. Discretionary managers bring judgment and conviction, yet cognitive bias and capacity constraints limit their reach.
Our quantamental approach resolves this tension at the architecture level. The AI engine identifies statistical probability across 200+ alpha factors. The fundamental overlay validates structural reality through experienced human analysis. Every trade is a hypothesis tested by machines and confirmed by judgment before capital is committed.
The portfolio is not a collection of good ideas. It is a mathematically optimised structure where position sizing, sector exposure, and risk contribution are calculated outputs governed by pre-committed limits, regime-aware recalibration, and independent risk oversight with the authority to override the portfolio manager.
Long/short equity driven by quantitative signals across developed and emerging markets. Factor-neutral portfolios isolate alpha from beta, trading 2,000+ liquid instruments with multi-factor signals spanning value, momentum, quality, and sentiment.
Proprietary strategies built on non-traditional data unavailable to conventional funds. NLP on earnings calls and filings, satellite imagery, web-scraped transaction data, and sentiment feeds drive machine learning models designed to anticipate market moves ahead of consensus.
Systematic macro across equity indices, foreign exchange, rates, and commodities. Adaptive models respond to regime changes, volatility spikes, and correlation breaks. Diversification across asset classes reduces drawdowns while preserving return potential.
We own the entire stack from data collection through execution. No third-party dependencies. Complete control.
150+ alternative and traditional data sources processed in real-time. Proprietary connectors normalise disparate formats and maintain data quality through automated validation.
Machine learning models trained on decades of market data identify alpha signals across multiple time horizons. Ensemble methods combine hundreds of factors with continuous retraining cycles.
Systematic optimisation maximises expected returns subject to risk constraints. Dynamic position sizing adjusts for regimes and liquidity, ensuring stability and responsiveness.
Real-time tracking of VaR, expected shortfall, correlation matrices, and concentration. Stress testing against historical crises and hypothetical scenarios at every level.
Sub-minute precision across 2,000+ instruments globally. Minimised market impact through intelligent order routing and timing optimisation across venues and geographies.
Continuous development of new alpha signals, stress-testing of assumptions, and enhancement of risk frameworks. Our quant team publishes internal research driving ongoing improvement.
Risk management is not a compliance function at Cavari Capital. It is the core architecture of our strategy. We embed risk controls at every layer, from individual positions through portfolio construction to firm-level aggregation.
Independent review of model methodology, third-party validation of backtests, and continuous monitoring of operational risks ensure our infrastructure meets the standards institutional investors require.
Stop-loss limits and liquidity reserves
VaR, expected shortfall, correlation analysis
Drawdown monitoring and volatility targeting
Concentration risk and capital adequacy
Deal flow and market intelligence from our advisory division inform and improve our capital strategies, creating information advantages unavailable to standalone funds.
Built on Accreta, our entire stack is proprietary. We control data, models, execution, and risk systems with speed and reliability competitors cannot match.
NLP, satellite imagery, transaction data, and sentiment analysis provide signals that traditional funds cannot access, transformed into alpha through advanced machine learning.
Significant GP commitment ensures our interests are aligned with our investors. We succeed only when our LPs succeed over long time horizons.
We operate under strict transparency standards with quarterly reporting, detailed performance attribution, factor exposure analysis, and position-level risk metrics. Our investor relations team is accessible at all times to discuss portfolio developments, strategy, and outlook.