Remote Machine Learning Engineer - Energy Forecasting / Mid+ (part-time)

ForecastPower sp. z o.o. develops commercial forecasting models for energy data including prices, generation, and consumption, focusing on stability, repeatability, and quality in a production environment.

• Minimum 3 years of experience in machine learning or data science
• Experience in time series forecasting
• Very good knowledge of Python and packages pandas, numpy, scikit-learn
• Practical experience with regression models
• Knowledge of tree-based and boosting models (XGBoost, LightGBM)
• Feature engineering for time series
• Model validation and data leakage elimination
• Independence and responsibility for the forecasting process

Nice to have:

• Experience in energy sector / renewable energy / energy trading
• Weather data in forecasting models
• Quantile regression, probabilistic models, ensemble forecasting
• Experience deploying and maintaining stable models in production environment

• Develop commercial forecasting models for energy data: prices, generation, consumption
• Build and integrate forecasting models in a production environment
• Ensure stability, repeatability, and quality of forecasts
• Parameterize forecasting models based on time series
• Select variables and design features (feature engineering) such as lags, moving statistics, seasonality, calendar, weather
• Implement models including multiple regression, decision trees, and other approaches
• Validate models using backtesting
• Eliminate data leakage
• Integrate models with ETL via API using a dedicated Python package
• Build training pipelines and periodic model retraining
• Monitor model quality and stability