Investment Philosophy

Aeonaux Capital operates at the confluence of quantitative finance, machine learning, and institutional-grade portfolio management. Our philosophy rejects narrative-driven investing, focusing instead on identifying and systematically harvesting verifiable risk premia and market inefficiencies through a rigorous, data-centric approach. The objective function is singular: maximize risk-adjusted returns (Sharpe, Sortino ratios) while adhering to strict, predefined risk constraints and minimizing correlation to traditional asset classes.

The core process involves continuous cycles of research, development, validation, and deployment. Hypotheses, often derived from academic finance, market microstructure analysis, or behavioral finance principles, are subjected to intense empirical scrutiny using terabytes of historical and real-time market data. We employ a diverse arsenal of quantitative techniques, from advanced econometrics and time series analysis (e.g., co-integration, GARCH models) to sophisticated machine learning algorithms (e.g., gradient boosting, recurrent neural networks, reinforcement learning for execution optimization) tailored to specific signal generation or risk prediction tasks.

Strategy Focus: Volatility & Derivative Alpha

A primary domain for alpha extraction lies within the global derivatives markets, particularly focusing on volatility as an asset class. Our strategies delve deep into the volatility surface, seeking to capture premia associated with variance risk (VRP), skew risk (SRP), and term structure dynamics. This manifests in systematic trading of options, futures, variance swaps, and volatility indices (e.g., VIX). Example strategies include statistical arbitrage on volatility indices vs. underlying index options, relative value trades on implied vs. realized correlation, and model-based selling of systematically overpriced options (e.g., earnings events, macro announcements) while dynamically hedging all primary Greek exposures (Delta, Gamma, Vega).

Execution is paramount. We utilize proprietary algorithmic execution logic, often incorporating reinforcement learning agents trained to minimize slippage and market impact based on real-time order book dynamics. Low-latency infrastructure and direct market access are critical components for effectively implementing these high-frequency hedging and trading strategies.

Uncompromising Risk Management

We view risk management not as a constraint, but as an integral component of alpha generation. A robust, multi-layered risk framework permeates every aspect of our process:

  • Factor Neutrality: Employing statistical factor models (e.g., Barra, Axioma) and optimization techniques to minimize exposure to unintended systematic risks (market beta, style factors).
  • Real-Time Risk Monitoring: Continuous calculation and monitoring of portfolio Value-at-Risk (VaR), Conditional VaR (CVaR), stress tests, and sensitivity analysis across multiple risk dimensions.
  • Automated Drawdown Controls: Hard and soft drawdown limits implemented at both the strategy and portfolio level, triggering automated position reduction or de-risking protocols.
  • Model Risk Management: Ongoing validation, parameter sensitivity testing, regime shift detection, and decay analysis to mitigate the risk of model failure or overfitting.
  • Operational & Counterparty Risk: Diversification across prime brokers, rigorous due diligence, and robust operational procedures (including cybersecurity) to minimize non-market risks.

This uncompromising focus on risk ensures portfolio resilience, protects investor capital, and provides the stability required for long-term compounding.

Performance & Risk Objectives*

> 2.9 Target Sharpe Ratio
< 5.0% Max Drawdown Limit
~65% Target Return (Ann.)
< 0.2 Target Beta (SPX)
Strict Tail Risk (CVaR) Control
High Strategy Capacity

*Targets based on model simulations & historical analysis, gross of fees. Not forecasts or guarantees. Investing involves significant risk. Past performance non-indicative of future results.

Methodology Stack

  • Modeling Approaches: Stochastic Calculus (Volatility Models), Advanced Time Series (Vector Autoregression, Cointegration), ML/AI (Gradient Boosting, RNNs, Transformers), Bayesian Methods.
  • Feature Engineering: Market Microstructure variables (Order Book Imbalance), Alternative Data integration, Cross-Asset Factor construction, Non-Linear Transformations.
  • Portfolio Construction: Hierarchical Risk Parity (HRP), Minimum Torsion Optimization, CVaR Minimization, Black-Litterman framework integration.
  • Validation Rigor: Combinatorial Cross-Validation, Stationary Bootstrapping, Reality Check (White's RC), Deflated Sharpe Ratio (DSR) calculation.

Technology & Infrastructure

  • Core Language: Primarily Python (scientific stack), C++/Rust for performance-critical execution/data processing modules.
  • Data Platform: Cloud-native architecture (AWS/GCP), time-series databases (TimescaleDB/InfluxDB), distributed compute frameworks (Ray/Dask).
  • Execution Systems: Custom low-latency gateways, FIX protocol integration, co-location facilities, sophisticated Transaction Cost Analysis (TCA).
  • Research Environment: Jupyter ecosystem, MLFlow/Weights & Biases for experiment tracking, proprietary simulation engines, robust CI/CD pipelines.