OmneFlux

AI Models

Sign In

Quant Model Center

5-category factor library · 6-stage AI quant pipeline · Full scientific metrics (IC/ICIR/Sharpe/Sortino/Calmar/Alpha) · Free daily optimization for users · $10,000/mo developer subscription for full build.

Quant Model Hub

Quant Model Center

Users get free daily log replay and one-per…

Users get free daily log replay and one-per-day factor optimization. Developers unlock AI-assisted strategy generation, portfolio construction, backtesting, and publishing with a $10,000/month subscription.

Open Studio (Log Analysis)

User Lane (Free)

Log replay → factor opt…

Log replay → factor optimization preview → confirm → apply in real time.

  • One free optimization per strategy / portfolio per day
  • Preview impact before any parameter update
  • Full IC / ICIR / Sharpe / Sortino / Alpha metrics suite

Developer Lane ($10,000/mo)

Subscription unlocks AI…

Subscription unlocks AI-assisted strategy build, portfolio assembly, full backtesting, and marketplace publishing.

  • AI-assisted single & multi-strategy portfolio generation
  • Full scientific backtest: Sharpe / Sortino / Calmar / Alpha / IC / ICIR
  • Publish to marketplace and earn profit share after threshold checks

Factor Library

Factor Library

Five factor categories covering the compl…

Five factor categories covering the complete quantitative research chain from signal generation to risk governance.

Momentum

Price Momentum

IC 0.03–0.12

12-1 month cross-sectional momentum score

Earnings Momentum

IC 0.04–0.15

EPS revision direction and magnitude

Short-Term Reversal

IC -0.02–0.08

1-month short-term price reversal

Quality

Earnings Stability

IC 0.02–0.10

Inverse EPS volatility over 5 years

Return on Equity

IC 0.03–0.11

Net income / shareholders' equity

Balance Sheet Quality

IC 0.02–0.09

Accruals ratio (low accruals = high quality)

Volatility

Realized Volatility

IC -0.04–0.06

20-day realized vol — low-vol premium

IV Skew

IC 0.01–0.08

Put/call IV spread — options market signal

Event-Driven

Earnings Surprise

IC 0.05–0.18

Actual EPS vs. consensus forecast deviation

News Sentiment

IC 0.02–0.12

NLP-scored sentiment from company news flow

Microstructure

Order Flow Imbalance

IC 0.03–0.14

Net buy/sell order flow imbalance ratio

Liquidity Score

IC 0.01–0.07

Amihud illiquidity ratio (lower = more liquid)

Quant Model Pipeline

Quant Model Pipeline

Data ingestion → feature engineering → si…

Data ingestion → feature engineering → signal generation → portfolio construction → risk management → execution, with AI modules at each stage.

1

Data Ingestion

OHLCV + alt-data + macro signals — real-time ingestion and cleaning

AI

Anomaly detection + auto imputation

2

Feature Engineering

Factor normalization, cross-sectional z-scoring, sector neutralization

AI

Automated feature selection and combination

3

Signal Generation

Multi-model ensemble inference — ranked opportunity queue

AI

RL dynamic weighting + ensemble voting

4

Portfolio Construction

Mean-variance / risk parity / Black-Litterman — selectable modes

AI

AI auto-selects optimal optimizer regime

5

Risk Management

VaR / CVaR computation, drawdown caps, position limit enforcement

AI

Real-time risk factor monitoring + auto-hedge trigger

6

Execution

TWAP / VWAP smart routing, market impact and slippage modeling

AI

Optimal execution timing prediction

Scientific Evaluation Framework

Scientific Evaluation Framework

Industrial-grade performance metrics powe…

Industrial-grade performance metrics powering model training and selection.

Metric

Formula

Good Threshold

Usage

Information Coefficient (IC)

Spearman(Factor, Fwd Return)

> 0.05 credible, > 0.10 significant

Core factor efficacy signal

IC Information Ratio (ICIR)

Mean(IC) / Std(IC)

> 0.5 stable, > 1.0 high-quality

Signal consistency over time

Sharpe Ratio

(Rp − Rf) / σ

> 1.0 good, > 2.0 excellent

Risk-adjusted return assessment

Sortino Ratio

(Rp − Rf) / σ_downside

> 1.2 good, > 2.0 excellent

Penalizes only downside vol — better for asymmetric risk prefs

Calmar Ratio

Ann. Return / Max Drawdown

> 1.0 good, > 2.0 excellent

Annualized return per unit of drawdown risk

Alpha (α)

Rp − (Rf + β·(Rm−Rf))

> 0% beats benchmark, higher is better

Excess return above benchmark — skill attribution

Beta (β)

Cov(Rp, Rm) / Var(Rm)

0.3–0.8 low-vol, ≈1 = market-neutral

Portfolio sensitivity to systematic market risk

Max Drawdown

max(Peak − Trough) / Peak

< 5% excellent, < 15% acceptable

Worst-case capital loss scenario for the portfolio

AI Training Directions

AI Training Directions

Continuous multi-paradigm AI training pip…

Continuous multi-paradigm AI training pipelines anchored to the factor library and scientific metrics framework.

Supervised Factor Scoring

Train gradient-boosting and deep-learning models on N-day forward returns to score and rank factor sets for relative return prediction.

RL Portfolio Rebalancing

Use Sharpe / Calmar as reward signal to train a policy network that learns optimal rebalancing timing and position sizing across market regimes.

Transfer Learning Cross-Market

Pre-train on liquid markets (e.g., US equities) and transfer to data-scarce venues (prediction markets, small-caps) to boost signal quality.

Meta-Learning Regime Adaptation

Train meta-models that adapt within hours of market regime shifts, addressing the stale-model problem in traditional quant approaches.

NLP News & Sentiment Signals

LLM-based real-time sentiment scoring of company announcements, analyst reports, and news flow — feeds high-frequency event factors into the signal layer.

Automated Factor Mining

Genetic programming and symbolic regression to auto-discover new alpha factor expressions, filtered by IC / ICIR backtests with overfitting penalties.

Capability Matrix

Capability Matrix

One interface, role-gated capability dept…

One interface, role-gated capability depth and workflow guardrails.

Capability

User (Free)

Developer ($10,000/mo)

Copy-Trade Log Replay

Full replay with Agent event stream

Includes attribution and comparison

Factor Optimization

1 free optimization per strategy/portfolio per day

Unlimited, all-factor-dimension tuning

Impact Preview

Preview return / drawdown delta first

Full Sharpe / Alpha / IC impact preview

Strategy Blueprint Generation

AI-assisted single & portfolio generation

Scientific Backtest Metrics

Full Sharpe / Sortino / Calmar / Alpha / Beta / IC / ICIR suite

Marketplace Publishing

Personal use only

Publish to marketplace after eligibility check

Smooth Workflows

User: Log Replay + Daily Optimization

1Open Strategy Studio and select the 'Copy-Trade Log Analysis' tab.
2Choose a strategy and analysis window, then click 'Analyze Logs'.
3Review full metrics: P&L, Sharpe, Alpha, IC, ICIR, and trade attribution.
4Trigger free daily optimization — preview factor deltas, confirm to apply.

Smooth Workflows

Developer: Build, Backtest, Publish

1Activate Developer Model Subscription ($10,000/mo) and switch to Developer mode.
2Select asset class, set strategy objective, and click 'Generate Blueprint'.
3Run backtest — review full Sharpe / Sortino / Calmar / Alpha / IC / ICIR suite.
4Use AI Tune to optimize parameters, confirm the preview, and re-run backtest.
5Submit to marketplace once backtest thresholds and publish eligibility are met.

footer.copyright

Risk notice: Strategy trading involves loss risk, and past performance does not guarantee future results. This platform does not provide investment advice.

Business & Support:[email protected]

Release: 2026.04