Investor OS

Autonomous AI Trading System.

OVERVIEW

Investor OS is a production-ready autonomous AI trading system built entirely in Rust. It combines a Hierarchical Reasoning Model (HRM) with real-time market regime detection, Kelly Criterion position sizing, and institutional-grade execution — all with sub-millisecond inference latency.

From signal generation to trade execution, every decision is auditable, every position is risk-managed, and every model inference is tracked with full lineage.

0.3ms HRM Latency p50 inference time
3,200/s Throughput Requests per second
290 Rust Modules Native performance
385 Tests 100% passing
CORE CAPABILITIES

HRM AI Model

Multi-factor conviction scoring: PEGY ratio, insider sentiment, social sentiment, VIX, market regime → Buy/Sell/Hold with confidence percentage.

Market Regime Detection

AI classification of market conditions: Trending, Ranging, Volatile, Crisis, Recovery. Automatic strategy adaptation.

Execution Engines

TWAP (Time-Weighted), VWAP (Volume-Weighted), Market orders. Kelly Criterion optimal position sizing.

Risk Management

Drawdown protection (daily/weekly/monthly), position limits, VaR calculation, correlation monitoring, emergency kill switch.

Portfolio Analytics

Real-time P&L tracking, performance attribution, tax loss harvesting, wash sale detection, backtesting.

INSTITUTIONAL SECURITY

5 Clearance Levels

Public → TopSecret role-based access control

HSM Encryption

AES-256-GCM with automatic key rotation

5 2FA Methods

TOTP, HOTP, WebAuthn, SMS, Email

Fireblocks

MPC multi-sig real crypto custody

Audit Trails

Immutable decision logging

Anti-Spoofing

Nonce replay, signatures, fingerprinting

TECHNOLOGY
Rust + AxumBackend
Burn MLNative Rust ML
Next.jsDashboard
TimescaleDBTime-series data
EU AI ACT GOVERNANCE
Art. 6

Risk Classification

Automatic model risk assessment. Four levels: critical, high, medium, low.

Art. 9

Model Registry

Every AI system registered with risk level, FRIA flags, and oversight requirements.

Art. 11

FRIA Generator

Fundamental Rights Impact Assessment auto-generated for high-risk models.

Art. 13

Transparency

Model cards, decision explainability, user notifications of AI-driven decisions.

Art. 14

Human Oversight

Kill switch, approval workflows, role-based access with full audit trail.

Art. 15

Drift Monitoring

Real-time performance drift detection with baseline tracking and alerts.

Ready to deploy intelligent trading?

Native Rust. Sub-millisecond latency. Fireblocks custody. EU AI Act compliant.