This list contains only the countries for which job offers have been published in the selected language (e.g., in the French version, only job offers written in French are displayed, and in the English version, only those in English).
We are seeking a Machine Learning & AI Developer to design, build, and deploy data‑driven and AI solutions that enhance catastrophe modelling, with an emphasis on flood hazard and risk. You will work with hazard and vulnerability modellers, software engineers, and other experts to prototype and industrialise novel approaches that improve the accuracy, scalability, and resolution of Impact Forecasting’s models. While the primary focus is flood, you will also support cross‑peril initiatives (e.g., wildfire, terrorism) and internal AI tools that accelerate research and development. A key initial responsibility will be contributing to an ML/AI‑based flood map quality assessment pipeline for new global flood models, replacing a manual QA process.
Job Responsibility
Develop and implement ML/AI solutions to improve catastrophe models, focusing on flood hazard and risk
Build global flood risk quality classification models using geospatial, hydrological, and exposure datasets
Contribute to the design, development, and maintenance of the Impact Forecasting Knowledge Hub (AI chatbot), including data pipelines, model selection, and monitoring
Research and implement precipitation downscaling methods (e.g., CNNs, U‑Nets, diffusion, generative models) to improve spatial and temporal resolution of inputs
Design, run, and document proof‑of‑concept (PoC) studies, from problem framing to evaluation and recommendations
Requirements
Proven experience (industry or advanced academic) in machine learning and/or applied AI, ideally on structured, time series, or geospatial data
Strong Python skills with common ML/AI and data science libraries (e.g., NumPy, pandas, scikit‑learn, PyTorch and/or TensorFlow)
Hands‑on experience with deep learning architectures (e.g., CNNs)
exposure to image‑like or gridded data architectures (e.g., U‑Net) is a plus
Solid understanding of ML fundamentals (training, validation, overfitting, metrics, hyperparameter tuning)
Experience with large, complex datasets, including data cleaning, feature engineering, and efficient processing
Secondary education or professional training with a school-leaving exam
English (Proficient)
Nice to have
Experience with geospatial data and tools (e.g., xarray, rasterio, geopandas, GDAL) or climate / environmental / remote sensing datasets
Knowledge of hydrology, meteorology, climate science, or natural catastrophe modelling
Familiarity with generative models (e.g., diffusion models, GANs) for downscaling or image‑to‑image tasks
Experience building or integrating chatbots or retrieval‑augmented generation (RAG) systems using LLMs
Exposure to MLOps / productionisation or cloud environments (preferably Azure/Databricks)