ML Cluster Alumni Spotlights: Jakob Schlör Shapes the Next Generation of Weather Forecasting

Jakob Schlör is currently working as a machine learning scientist at the European Centre for Medium Range Weather Forecasting, an institution where he developed his interest during his PhD by utilizing their globally recognized data products and networking with researchers from conferences there. His primary role involves creating sub-seasonal machine learning models that forecast weather conditions beyond the typical ten-day window used in standard predictions, extending up to six weeks into the future. These forecasts focus on weekly averages over periods of three or four weeks, providing critical insights for sectors like agriculture regarding planting schedules and drought risks, as well as assisting electricity providers in managing renewable energy production and demand planning. The models Jakob develops are trained on forty years of historical data, primarily from satellites, representing a significant leap forward compared to traditional numerical weather prediction methods that have been refined over the same period. Remarkably, these new machine learning approaches, which were developed within just three years, now outperform decades of conventional model development in accuracy and efficiency. This technological advancement allows for more reliable long-term outlooks that are essential for strategic planning in industries dependent on stable climate conditions, marking a pivotal shift from short-term weather alerts to broader seasonal forecasting capabilities. Beyond the technical achievements, Jakob highlights the profound value of the collaborative environment fostered during his doctoral studies, which connected him with peers and colleagues who remain friends today despite working across diverse fields united by machine learning. He emphasizes that this interdisciplinary exposure provided him with a comprehensive overview of various domains within the field, resources he frequently references in his current work. However, he also acknowledges the mental challenges inherent in PhD research, where projects often deviate from plans and researchers can feel isolated when comparing their progress to others who seem more successful in publishing papers. To navigate these difficulties, Jakob advises aspiring researchers not to hesitate in seeking help or talking to peers, as most colleagues are likely experiencing similar struggles with loneliness and uncertainty. He suggests that forming joint projects is a powerful strategy for maintaining motivation and enriching the research experience, noting that it does not matter whether an initiative is individual or collaborative since no one ultimately cares about ownership but rather about collective success. By fostering open communication and teamwork, researchers can create a supportive community where shared challenges are addressed together, making the journey of scientific discovery more manageable and rewarding for everyone involved.