Role and objective
The Machine Learning Expert will be responsible for developing a comprehensive machine learning impact based forecasting (IBF) system to support decision-making and enhance resilience in Eastern Africa. The IBF system will leverage socio-economic, meteorological, and impact data to provide forecasts that emphasize the consequences of weather/climate events. This system will be designed to provide actionable insights, improve resource allocation, support proactive financial planning, and enhance preparedness across key sectors.
Main tasks and responsibilities
Data Collection & Analysis:
Conduct an assessment to identify gaps and limitations in current datasets related to impact-based forecasting.
Collect and pre-process datasets, including historical weather data, agricultural yield data, flood records, and socio-economic indicators.
Model Development:
Develop and implement machine learning models for predicting potential impacts on agriculture, water resources, and vulnerable populations using climate and socio-economic data.
Test and refine the models to account for varying impacts across different geographical regions.
System Integration & Automation:
Integrate the developed machine learning models into an automated system for regular data updates and real-time impact forecasting.
Ensure seamless connectivity between various databases and forecasting tools.
Validation and Prototype Development:
Validate model performance using historical weather events and impact records.
Develop a prototype system for yield forecasting over key agricultural zones in Kenya (or other pilot regions as per the project plan).
Capacity Building & Training:
Organise workshops and training sessions to enhance stakeholders' understanding of the system and its outcomes.
Prepare technical documentation and guidelines for the system's usage.
Qualifications
Advanced degree in Computer Science, Data Science, Statistics, Geo-Informatics, Meteorology or Climate Sciences or equivalent
At least five years of experience, two of which were in developing Machine Learning methods to solve specific problems, with particular interest towards scientific applications
Experience developing, debugging and applying models using modern deep learning frameworks
Proficiency in scripting and programming languages preferably Python and/or R programmes; experience analyzing big data
Experience with ML systems using frameworks such as Scikit-learn and Tensorflow
Good understanding of Machine Learning concepts and methods when to apply them and how to effectively implement them using the available machine learning packages is key
Familiarity with Git, Docker, AWS or equivalent
Experience and understanding of statistical and geo-statistical techniques and how to apply them in various contexts including climate applications.
Ability to describe findings and the way techniques work to audiences, both technical and non-technical and visualisation of the results using various tools.
Experience in integrating and visualising products and outputs on web-based platforms. An understanding of various web development techniques to make products available on frontend systems from backend procedures is desirable.
Experience/interest in climate data, climate science and disaster risk