AI-native Analytical Studio · Prague

Quantitative intelligence for
global energy markets

EdgeTech is an independent analytical research studio combining machine learning, quantitative modelling and alternative data to generate actionable insights across global power, gas, LNG and carbon markets.

Get in touch
OHLCV · Bollinger MA20 MA50
PWRGAS CO2OIL LNGFX PWRGAS CO2OIL LNGFX −1 0 +1 cross-asset ρ
PC1 PC2 A B C
Σ quantitative
ATTENTION WEIGHTS t-6 t-5 t-4 t-3 t-2 t-1 t t-6t-5 t-4t-3 t-2t-1 t head 1 head 2 head 3 self-attention / transformer
σ (risk) return ★ efficient frontier
Power €/MWh TTF Gas €/MWh EUA CO₂ €/t Brent $/bbl
AI
Machine learning at the core of every model
100%
Independent — no conflicts, no sell-side bias
Global
Power · Gas · LNG · Carbon across EU & world markets
INPUT LSTM ×2 ATTN DENSE OUT x₁x₂xₙ ∫∫ ∫∫ ααα ŷ₁ŷ₂ skip connection LSTM · ATTENTION · DENSE

Who we are

AI-native research, built for global markets

EdgeTech was founded with a single ambition: to bring institutional-grade quantitative analysis to energy and commodity markets — independently, without conflicts of interest and with the full power of modern AI.

Our proprietary data infrastructure connects macroeconomic signals, fundamental market data and alternative datasets. Machine learning models process these inputs in real time, delivering high-resolution insights to clients across Europe and globally.

Contact us

Our methodology

Quant research meets modern data science

We combine classical econometric modelling with state-of-the-art machine learning — LLMs for qualitative signal extraction, gradient boosting for price forecasting, and neural networks for pattern recognition across multi-market datasets.

Contact us
FEATURE IMPORTANCE gas_fwd_1m implied_vol temperature renewables_% CO2_price storage_lvl oil_brent 0.21 0.17 0.14 0.12 0.10 0.07 0.05 SHAP VALUES 0 gas_fwd_1m implied_vol temperature renewables_% CO2_price storage_lvl oil_brent low high PREDICTED vs ACTUAL predicted actual RMSE = 3.24 R² = 0.93 ROC CURVE FPR TPR 0 0.5 1 0 0.5 1 AUC = 0.89 AUC = 0.84 GBM · XGBoost · GRADIENT BOOSTING DIAGNOSTICS

What unites us

The full cycle — from data to decision

  • AI-native platform — machine learning embedded across the entire research stack
  • Quantitative signal generation from structured and unstructured data sources across multiple asset classes
  • Systematic approach covering the full cycle — raw data ingestion, feature engineering, model training, back-testing and live signal delivery
  • Global market coverage: European and international power, gas, LNG, carbon and renewable energy certificates
Contact us
RAW DATA feeds · APIs DATA LAKE clean · store FEATURE STORE engineer · scale MODEL TRAIN GBM · LSTM · RF BACKTEST ───────────── LIVE SIGNAL validate · deploy retrain loop BACKTEST CUMULATIVE P&L +40% +20% 0% −10% 2022 Q2 2023 Q2 2024 Sharpe 1.84 Max DD −8.3% CAGR +22% Hit Rate 57.4% IC 0.47 Calmar 2.65 rolling 60d Sharpe 1.0 DATA → FEATURES → MODEL → SIGNAL → ALPHA