Approach

One model. Many expressions. Fixed discipline.

Strategy, risk, and technology are designed as a single system: signals come from the model, instruments are chosen to fit the signal, and risk rules are committed before capital moves.

Strategy

An instrument‑agnostic core.

The firm maintains a library of systematic strategies spanning directional, volatility, and hedged-income families, all driven by a single quantitative core. The core is instrument-agnostic by design: it reads the structure of price series, not the idiosyncrasies of any product, so a time series is a time series whether it belongs to a stock, a fund, or an index. The same signal can then be expressed as an equity position, an ETF, or a defined-risk derivative structure, whichever carries the cleanest form.

Derivatives are used deliberately: to define risk in advance, to shape payoff asymmetry, and to monetize structure that linear instruments cannot reach.

Risk

Risk is a budget, decided in advance.

PRE‑COMMITTED LIMITS

Maximum loss per trading cycle is fixed before the cycle opens. Position sizing derives from that budget, never from conviction.

DEFINED‑RISK STRUCTURES

Where derivatives are used, structures are chosen so worst-case outcomes are known at entry, not discovered at exit.

LIVE-MARKET VALIDATION

No strategy earns a place in the catalog on backtests alone. Each one must pass through backtest, paper, and live stages with the firm's own capital at risk at every step.

Technology

From raw series to executed risk.

I · DATA

Market data is treated as the raw material: cleaned, structured, and versioned so research is reproducible.

II · SIGNAL

The firm's proprietary time-series foundation model and machine-learning stack generate and rank candidate signals.

III · STRUCTURE

Signals are mapped to instruments and, where appropriate, to defined-risk derivative structures sized against the risk budget.

IV · EXECUTION

Rules govern entry, management, and exit. Agentic execution by autonomous agents, inside the same discipline, is the active research frontier.

Research Program

Built, not bought.

The firm’s edge is engineered in-house. The time-series foundation model anchors the research stack; around it sit the strategy library, the risk engine, and a growing body of work on agentic AI for research and execution.

We publish no forecasts and sell no signals. The research exists for one purpose: to trade the firm’s own capital well.