Devops & Machine Learning Expert

Specific Responsibilities

Under the overall guidance of the DRM Program Head, the DevOps & Machine Learning Expert will undertake the following tasks:

• The hydrological forecasting chains in use at ICPAC running operationally in an automated forecast cycle, with new product releases deployed without interruption to service.
• Harmonised regional hydrological reference and observation datasets, with documented quality control and exchange interfaces for national hydrological services.
• Operational STAC API and analysis infrastructure for hazard, exposure and impact data, with reliable processing pipelines across all integrated models.
• East Africa Flood Watch in public operation at functional parity with Drought Watch, with its warnings and products integrated into the East Africa Hazard Watch Portal.
• Automated data processing and reporting workflows, including impact-based forecast bulletin generation with language-model-assisted drafting under documented evaluation and guardrails.
• Technical support provided to Member States, NMHSs and regional partners on IGAD flood monitoring activities, with a record of requests addressed.
• Calibrated post-processing and multi-model ensemble combination in operations, supported by a methodology note.
• Bayesian Network risk model and inference service deployed and integrated with the risk monitoring application.
• Event-based climate storylines contributed to the drought and flood event catalogue.
• Forecast verification framework with routine skill reporting, and documented governance for every operational model.
• Reproducible multi-cloud infrastructure as code with ecFlow and Prefect scheduling, CI/CD and GitOps delivery, observability in production, and a tested disaster-recovery procedure.
• Complete technical documentation, deployment guides, JupyterHub tutorials, training materials and stakeholder hand-over.
• Performs such other duties as may be assigned from time to time.

Key Educational Qualifications and Professional Experience

Academic Qualification

• University degree in Computer Science, Geo-Informatics, Hydroinformatics, Computer Engineering, Data Science, Software Engineering, Information Technology or other relevant field; an advanced degree is an advantage.
• Candidates should demonstrate their qualifications and proficiency in web application development and geo-application development (provide links to at least 2 samples of previous work and/or Github code)

Professional work experience

• Minimum of four (4) years of relevant experience in geo-applications design and development.
• Proficiency in web application development and geo-application development, demonstrated through a portfolio of developed products (at least two samples of previous work and / or GitHub code).
• Demonstrated experience operating production cloud infrastructure under daily operational deadlines, and deploying machine learning models operationally rather than in research settings.
• Experience supporting national institutions in an operational early warning context is desirable.
• Scripting and automation of geoprocessing and large data workflows, especially in Python; sound knowledge of SQL and PostGIS.
• Web technologies (HTML, CSS, JavaScript) and production-ready geospatial web applications using Node.js, React, Mapbox GL, Leaflet, GeoServer, MapServer and GDAL; REST API development and microservices architecture; experience with Go is an advantage.
• OGC geospatial standards including WMS, WFS, WCS, WPS and Simple Features for SQL; handling and analysis of Earth Observation data in a range of formats.
• STAC API and PySTAC; workflow management with ecFlow and Prefect; xarray, dask and the numpy ecosystem; rasterio and geopandas; GRIB2, Zarr, COG, VirtualiZarr, Icechunk and kerchunk; PostgreSQL/PostGIS and TimescaleDB.
• Hydroinformatics: operationalisation of rainfall-runoff, hydrodynamic and rapid inundation forecasting chains; forcing preparation; catchment, river network and terrain data management; hydrometric and remotely sensed observation handling; and hydrological data standards and exchange.
• Container orchestration (Docker, Kubernetes, Helm) and GitOps delivery (ArgoCD); multi-cloud computing on Google Cloud Platform and Amazon Web Services with Terraform, Coiled and CI/CD pipelines; CUDA and GPU environment management; observability, incident response, cloud security and cost management.
• Impact-based forecasting systems and climate modelling workflows; ensemble post-processing, calibration and downscaling; Bayesian networks; forecast verification; MLOps practice; LLM integration, evaluation and guardrails; training and capacity development.

Essential Skills and Competencies Required

• Self-driven, result-oriented, problem solver
• Teamwork
• Communication
• Continuous improvement and knowledge sharing