Data Scientist

This is a fully remote position, open to applicants in Germany.

📋 Description

• Assist in the design and execution of CLEAR’s data architecture.

• Collaborate with developers and data engineers to establish the necessary environment for training and refining models, including data ingestion pipelines, compute and GPU resources, experiment tracking, and MLOps tools, actively influencing the setup instead of passively awaiting its provision.

• Create machine learning and deep learning models that amalgamate various data streams to identify early signs of humanitarian crises, integrating earth observation and satellite imagery, conflict event data, climate monitoring, economic indicators, and population movement trends into cohesive risk assessment frameworks.

• Develop computer vision and remote sensing models utilizing satellite and aerial imagery (such as flood extent mapping, building and settlement detection, displacement site monitoring, infrastructure damage assessment, and land cover change detection) and combine these earth observation results with non-imagery data sources.

• Aid in the development of automated alert systems that recognize emerging crises, calibrating predictive algorithms for various crisis types.

• Formulate ensemble modeling approaches that merge traditional statistical techniques with advanced AI methodologies.

• Refine and adapt foundational models and large language models for humanitarian applications such as document triage, multilingual report analysis, and situation summarization.

• Investigate methods for adapting models to changing crisis conditions through reinforcement learning systems.

• Execute impact-based forecasting systems that convert meteorological, conflict, and economic forecasts into specific humanitarian outcomes, such as displacement volumes, food insecurity levels, and estimates of infrastructure damage.

• Construct decision trees and recommendation systems that assist field personnel in systematic needs assessment processes informed by predictive analytics and historical response data.

• Develop automated reporting systems and interactive dashboards that provide field teams and leadership with access to real-time data for prompt response activities.

⛳️ Requirements

• Advanced degree in Data Science, Statistics, Computer Science, Physics, Engineering, Economics, or a related quantitative field, with at least 5 years of professional experience in applied data science.

• Proven experience in designing, training, and fine-tuning deep learning models (e.g., CNNs, recurrent/sequence models, and transformers), including structuring training runs, managing compute resources, and diagnosing and enhancing model performance.

• Advanced proficiency in Python for statistical analysis, machine learning, and data manipulation, with experience in key libraries such as pandas, NumPy, scikit-learn, TensorFlow, PyTorch, and Keras.

• Strong SQL capabilities for database management and complex query optimization.

• Practical experience implementing both supervised and unsupervised learning algorithms, including regression models, classification techniques, clustering methods, and time series analysis.

• A demonstrated history of proactively sourcing and engineering data (finding, negotiating access to, cleaning, and where necessary, generating data).

• Experience in designing and implementing ETL pipelines for processing datasets from various sources, with skills in data cleaning, transformation, and quality assurance at scale.

• Ability to create automated reporting systems and dashboards utilizing tools like Tableau, Power BI, or similar platforms.

• Experience working with large, disorganized, real-world datasets.

• Understanding of model deployment and MLOps practices, and comfort collaborating with engineers to provision the necessary infrastructure for models.

• Fundamental skills with version control software and collaborative development.

• Fluency in written and spoken English; proficiency in other languages is advantageous.

• Practical experience in low-data or data-scarce environments, including transfer learning, few-shot learning, data augmentation, and synthetic data generation to mitigate limited training data challenges.

• Experience in implementing and fine-tuning large language models (LLMs) for applied tasks.

• Familiarity with natural language processing techniques for analyzing reports, social media, or news data pertinent to crisis monitoring.

• Experience utilizing GenAI for automated analysis of large volumes of documents, including extracting key themes, sentiment analysis, and identifying emerging trends across diverse contexts and languages.

• Understanding of ensemble methods and explainable AI techniques for transparent decision-making.

• Experience with earth observation and satellite imagery; including optical (e.g., Sentinel-2, Landsat) and radar/SAR (e.g., Sentinel-1), alongside commercial high-resolution sources, and geospatial tools such as Google Earth Engine, rasterio/GDAL, xarray, and geopandas.

• Applying computer vision and deep learning to imagery (semantic segmentation, object detection, change detection) for humanitarian tasks such as flood extent mapping, damage assessment, settlement and displacement-site detection, infrastructure monitoring, and population estimation is a strong asset.

• Experience with multimodal data fusion.

🏝️ Benefits

• Duty station: Remote (Germany, France, UK, or Belgium)

• Contract: Fixed term (2 years)

• Travel: Up to 10%

• NRC is an equal opportunities employer. We are dedicated to diversity without distinction to age, gender, religion, ethnicity, nationality, and physical ability.

• We encourage innovative thinking, welcome new ideas, and empower all employees at every level. You will have numerous opportunities to voice your thoughts and take initiative.