How an AI hedge fund decides — and when it refuses to.

The bedrock philosophy

Every layer of the system reads a shared written doctrine before it acts. The doctrine is short and load-bearing:

1. Survival before return. A blown account cannot compound. Capital preservation is the prerequisite for every other goal.
2. Uncertainty equals WAIT. The default action is not action. Confidence below threshold means hold, not size down.
3. The market is the boss. Stance follows reality. The risk posture — from active engagement to full lockdown — is not a knob the operator sets; it is the system's read of conditions.
4. Mistakes are journaled, not punished. Every miss writes to the audit trail. Tomorrow gets rewritten from what failed today.
5. Emergency is supreme. An independent safety system can halt everything, overriding every other layer. Failure is fail-closed.

The organization

iQntX is structured like a real fund: specialized layers, each owning one kind of decision. No layer owns the whole trade.

• Strategy layer — Sets stance, holds the weekend, declares code-red, responds to broker margin events.
• Research layer — Reads geopolitics, central banks, and macro flow. Forms an independent view of the market regime.
• Risk layer — Layered safety gates and exposure controls. The veto layer.
• Strategy bank — A catalog of setups, switched by regime. Confluence, backtest, mean-reversion, and pattern logic among them.
• Execution layer — Order management, coordination, and pre-trade verification.
• Intelligence layer — News and calendar ingestion, journaling, self-optimization, and periodic reporting.

The decision lifecycle

Every layer runs the same structured loop on every wake-up: read the doctrine, read the active context, read the active stance, read inbound messages, form a hypothesis, consult its playbook, emit a proposal, pass pre-trade verification, journal the result. No layer skips a step.

Layered safety gates

Every decision passes multiple independent gates before reaching the EA:

1. Risk gate — Stance + size + drawdown headroom + correlation exposure must all be within bounds.
2. Pre-trade verification — The inputs that justified the decision must still be true at the moment of signing.
3. Independent confirmation — A second, blind check must reach the same conclusion.

A trade that any gate rejects becomes a journal entry, not a position.

The emergency reflex

An independent safety system runs outside the agent graph. It monitors drawdown, margin utilization, broker connectivity, and stop-loss invariants. If any invariant breaks, it issues an automatic halt that closes positions, locks the EA, and pages the operator. No other layer can countermand a safety halt.

Compounding intelligence

The system continuously reviews the day's journal, scores what worked, retires what didn't, and queues experiments. It gets smarter without changing operators. The audit trail is human-readable — you can read why a setup was retired six months from now.