Loading
eTE AI · LabThe intelligence
The intelligence
behind every trade.
eTE AI is our proprietary machine learning engine. It processes tick data, sentiment, microstructure and on-chain data in real time: all accessible via API, platform or conversational assistant. It doesn't replace the trader: it amplifies them.
- +34%Directional accuracy in 5Y backtest vs ARIMA baseline
- 8msAverage engine decision latency
- 20K+Strategies backtested monthly by users
- 200+Real-time signals processed per second
- 12Of tick training data
- 99.9%Guaranteed AI engine uptime
Six modules. One platform.
Enable the ones you want from your Dashboard. Each works independently or combined with the others.
- ML
Predictive models
Neural networks trained on 12+ years of market data · ensemble learning over 200+ real-time signals.
- NLP
Live sentiment
NLP across 50K+ daily sources (news and social media) quantified to actionable metrics in milliseconds.
- Risk
Risk co-pilot
The engine evaluates each order against your portfolio, correlations and aggregate exposure before executing.
- Copilot
eTE Copilot
Conversational assistant inside the platform: 'analyze my USD exposure', 'what happened with XAU today'.
- Backtest
Strategy backtesting AI-assisted
Describí una estrategia en lenguaje natural, eTE AI la traduce a código, la backtestea y entrega métricas + análisis de overfit.
- Auto
Auto-execution con guardrails
Para cuentas Platino: el sistema ejecuta autónomamente dentro de parámetros que vos defines (max drawdown, whitelist, sizing max).
Five stages. Sub-40 ms.
The journey of each signal from raw data to execution (or alerting the trader).
- 01
Data ingestion
Tick data de 25+ proveedores Tier-1 + news feeds + social + on-chain crypto + macro releases. ~2.4 TB diarios.
- 02
Feature engineering
Pipeline propietario que extrae 200+ features por instrumento: volatilidad realizada, sentiment scores, microstructure, correlaciones cross-asset.
- 03
Ensemble de modelos
LSTM para series temporales + Transformers para sentiment + Gradient Boosting para regime detection. Salida ponderada por confianza.
- 04
Risk evaluation
Cada señal pasa por capa de gestión de riesgo que la contextualiza con tu portfolio y aplica filtros (DD, correlación, exposure caps).
- 05
Execution layer
Si la señal supera todos los filtros, se enruta al matching engine para ejecución sub-40ms. Si no, queda como señal informativa.
eTE AI is not a black box.
Technical transparency as a competitive differentiator.
Auditoría externa
Modelos auditados anualmente por Deloitte AI Practice. Reportes bajo NDA.
Performance pública
Backtests reproducibles publicados en GitHub: github.com/ete-research/ai-benchmarks
Explainability
Cada señal viene con su rationale: qué features contribuyeron, qué confianza tiene el modelo.
Activate eTE AI in your account.
Available on all plans from Silver. Guided setup in 5 minutes.