Data Scientist - Load & Renewable Generation Forecasting
Job Title: Senior Data Scientist
Location: Juno Beach, FL (Day one onsite)
Experience: 8 to 10 Year
Position Specific Description The IT Forecasting team is seeking a Senior Data Scientist to develop, enhance, and productionize forecasting solutions supporting Wind, Solar, and Load forecasting across NextEra Energy operations. The primary focus of this role will be Wind generation forecasting, while also providing technical support across Solar and Load forecasting initiatives. This is a hands-on role combining forecasting, meteorology, renewable-energy analytics, machine learning, data engineering, and production support. The successful candidate will independently develop and improve forecasting models from data exploration and feature engineering through validation, deployment, monitoring, and ongoing optimization. A key responsibility will be evaluating and optimizing the use of weather and Numerical Weather Prediction (NWP) providers. The candidate will compare provider accuracy, available variables, forecast horizons, geographic coverage, latency, reliability, and overlap between services to determine the most effective weather inputs for each forecasting application. This includes identifying where individual providers or weather features perform better under specific conditions and identifying opportunities to reduce unnecessary provider redundancy while maintaining forecast accuracy and resiliency. The ideal candidate combines strong statistical and machine-learning skills with practical understanding of the meteorological, physical, and operational factors affecting renewable generation and electric demand.
Key Responsibilities:
Forecasting & Model Development
Develop and enhance Wind, Solar, and Load forecasting models from research and backtesting through production implementation.
Build forecasting approaches using combinations of:
- Weather and NWP forecasts
- Historical generation and loadSCADA and operational telemetry
- Persistence and statistical baselines
- Physical relationships
- Gradient boosting and other machine-learning techniques
- Ensemble and probabilistic forecasting methods
- Perform feature engineering, model selection, optimization, and validation.
- Develop backtesting and benchmarking processes to evaluate forecast performance across sites, horizons, seasons, and operating conditions.
- Investigate forecast errors and determine whether they originate from weather, model behavior, data quality, or operational conditions.
- Improve existing and legacy forecasting models while identifying opportunities to simplify and modernize model architecture.
Weather Provider Analytics & Optimization:
Analyze multiple weather and NWP providers for accuracy, reliability, coverage, latency, available variables, forecast horizon, and operational value.
Compare provider performance by:
- Location/site
- Forecast horizon
- Weather variable
- SeasonWeather regime or event type
Identify the relative strengths and weaknesses of individual weather providers and determine where specific providers or variables may provide superior forecasting value.
Assess redundancy and overlap across weather services and support recommendations for an efficient core set of providers.
Develop and evaluate approaches for:
Weather-provider blending and ensembles
Provider weighting and selection
Bias correction
Feature selection
Provider fallback and substitution
Ensure correct alignment of weather forecast issue times, valid times, time zones, spatial resolution, measurement units, and forecast horizons.
Monitor weather-data quality and identify provider anomalies or degradation that could impact production forecasts.
Energy Forecasting Domain Knowledge
Apply relevant physical, meteorological, and operational knowledge across forecasting applications.
WindWind speed and direction
Hub-height adjustments
Vertical wind shear
Temperature, pressure, humidity, and air density
Turbine power curves
Cut-in, rated, and cut-out behavior
Wake and site effectsSCADA and generation telemetry
Curtailment, derates, outages, maintenance, and equipment availability
SolarIrradiance and cloud conditions
Ambient and module temperature
Inverter performance and availability
Tracker position and stowingPPC/plant-controller impacts
CurtailmentEquipment outages
Soiling and degradation
MaintenanceSite capability versus nameplate capacitySCADA and plant telemetry
LoadTemperature, humidity, heat index, and other weather drivers
Calendar and seasonal patterns
Lagged demand and weather relationships
Extreme-weather behavior
Regional and customer-segment differences
Peak-demand forecasting
The candidate should understand that observed generation and demand can be influenced by operational conditions in addition to weather and account for these factors when developing, training, and evaluating forecasting models.
Required Skills & Experience:
Advanced proficiency in Python and common data-science libraries such as pandas, NumPy, scikit-learn, statsmodels, LightGBM, XGBoost, or equivalent tools.
Demonstrated experience developing and deploying forecasting, time-series, or predictive models.
Strong knowledge of:
- Feature engineering
- Regression and time-series methods
- Gradient boosting
- Ensemble forecasting
- Probabilistic or quantile forecasting
- Model validation and backtesting
Experience integrating and analyzing weather APIs, NWP forecasts, SCADA/telemetry, databases, and external data feeds.
Ability to evaluate weather-provider performance and determine its impact on forecast accuracy.
Strong SQL and data-analysis skills.
Experience transitioning models from research and development into reliable production workflows.
Working knowledge of AWS services such as EC2, S3, Lambda, ECS, Step Functions, CloudWatch, and SageMaker, or comparable cloud technologies.
Experience with Git, version control, testing, and reproducible software-development practices.
Strong troubleshooting, analytical, communication, and documentation skills.
Ability to work independently while collaborating effectively with data scientists, data engineers, DevOps, operations, trading, and business stakeholders.
Model Performance & Production Support:
Monitor forecast performance using metrics such as MAE, RMSE, normalized error, bias, forecast skill, and probabilistic metrics where appropriate.
Benchmark internal forecasts against persistence, existing models, alternative weather providers, and third-party forecasts.
Develop monitoring and diagnostics for:
- Forecast degradation
- Data-quality issues
- Weather-provider outages or anomalies
- Model drift
- Operational impacts
- Implement appropriate logging, alerts, error handling, and recovery procedures.
- Support forecasting solutions designed to continue operating when individual data feeds or weather services become unavailable.
- Perform root-cause analysis of significant forecast misses and recommend corrective actions.
Production & Data Integration:
Develop data connectors and pipelines supporting weather, telemetry, historical generation, load, and operational data.
Support scalable batch and near-real-time forecasting workflows.
Maintain clear model configuration, versioning, dependencies, and documentation.
Collaborate with data engineering and DevOps teams to deploy and maintain forecasting applications.
Build dashboards and analytical tools using Streamlit, Plotly, or similar technologies to communicate forecast performance, model behavior, and operational impacts.