Senior Data Scientist

About Verdant Grid

Verdant Grid forecasts what wind, solar and battery assets will produce, and helps operators decide when to sell it. Our customers make trading decisions on our numbers every day, so being wrong is expensive and being confidently wrong is worse. The company is employee-owned, runs a four-day week, and has about 90 people spread across northern Europe.

What you'll do

• Improve our day-ahead and intraday generation forecasts, where a percentage point of accuracy translates directly into customer revenue.
• Work with genuinely difficult inputs: numerical weather prediction ensembles, sparse and unreliable telemetry, curtailment events, and assets that were recommissioned without anyone telling us.
• Own model evaluation properly, including calibration of uncertainty estimates rather than only point accuracy.
• Take models from notebook to production alongside our platform engineers, and stay responsible for them once they are running.
• Investigate forecast failures after the fact and explain to a customer, in plain language, what went wrong.
• Contribute to how we quantify and communicate uncertainty, which is increasingly what customers buy from us.

What we're looking for

• Five or more years of applied data science or quantitative modelling work, with production deployments you can describe in detail.
• Strong time series background: hierarchical forecasting, probabilistic methods, and a healthy scepticism about backtest results.
• Solid Python engineering practice. Your code should be testable by someone else without an archaeology expedition.
• Statistical rigour, particularly around uncertainty quantification and the difference between a model that fits and a model that generalises.
• The ability to explain a method to a trader who has no interest in the mathematics but every interest in whether to trust it.
• Located in Europe with at least four hours of overlap with Central European Time.

Nice to have

• Domain knowledge of power markets, dispatch, or renewable asset operations.
• Experience working with numerical weather prediction data or other large gridded datasets.
• Familiarity with running models under regulatory or audit scrutiny.

Benefits

• A four-day, 32-hour week at full pay. This is the default for everyone, not a negotiated exception.
• Six months of paid parental leave shared flexibly between parents.
• EUR 2,500 annual learning budget, plus conference travel.
• Employee ownership, with a share of company profits distributed annually.