Context information
The Rapid Response Mechanism (RRM) is an ECHO-funded project implemented in partnership with Danish Refugee Council. It provides timely, life-saving assistance to populations affected by new emergency events across Sudan. The RRM primarily responds to conflict-driven displacements, flooding, and health outbreaks through multi-sectoral interventions.
Given Sudan’s dynamic and rapidly evolving context, continuous monitoring is essential to ensure RRM partners are prepared and able to respond effectively. The RRM requires a strengthened system to anticipate, monitor, and respond to rapid-onset emergencies, including conflict-induced displacement, natural disasters, and public health crises. To address this, NRC will develop an enhanced Early Warning and Alert System (EWAS) that consolidates multiple information streams, monitors evolving risks, and provides rapid and actionable alerts to support response planning and coordination. This mechanism will provide timely, reliable, and actionable insights to support evidence-based decision-making, and enable an effective and efficient RRM.
2. Summary
The Data Science consultant will lead the development and implementation of the AI-powered components of the Early Warning and Alert System for the Sudan RRM. This role focuses on transforming diverse data sources into actionable insights through innovative AI approaches, enabling timely response to emergencies in Sudan.
3. Scope of Work
- Designing and implementing machine learning models for crisis detection and early warning
- Developing pattern recognition algorithms for identifying crisis signals from satellite imagery, media/social media, weather data, and conflict indicators
- Creating data processing pipelines to integrate multiple data sources (field reports, satellite imagery, social media, weather forecasts, conflict data)
- Implementing AI models that can function with limited training data in the Sudanese context
- Validating and testing models with humanitarian datasets specific to Sudan
- Comprehensively documenting model performance and limitations
- Building the initial dashboard and data visualization components
- Optimizing models for field deployment under Sudan's infrastructure constraints
4. Key Deliverables and Timeline
(see link)
5. Required Qualifications
Essential
· 5+ years experience in applied machine learning/data science
· Strong programming skills in Python and ML frameworks (TensorFlow/PyTorch)
· Experience with NLP for media monitoring and computer vision for satellite imagery analysis
· Knowledge of predictive modeling for crisis/conflict scenarios
· Experience working with limited and messy datasets typical in humanitarian contexts
· Master's degree or higher in Data Science, Computer Science, or related field
· Understanding of responsible AI principles and ethical considerations in crisis contexts
Preferred
· Experience with multimodal data fusion (combining satellite imagery, text, and structured data)
· Prior work in humanitarian or conflict settings, particularly in East Africa
· Experience with real-time alert systems
· Knowledge of GIS and spatial data analysis
· Familiarity with cloud computing infrastructure for model deployment
6. Application Requirements
Interested consultants should submit:
- A technical proposal (max 5 pages) outlining:
- Proposed methodology for implementing the three-phase approach
- Experience with similar AI-powered early warning systems
- Approach to handling data limitations in the Sudanese context
- CV/resume highlighting relevant experience
- Financial proposal
- Two examples of previous relevant work (dashboards, prediction models, or GitHub repositories)
7. Evaluation Criteria
Applications will be evaluated based on:
- Technical expertise in crisis-relevant AI applications (40%)
- Quality and feasibility of proposed methodology (30%)
- Understanding of Sudan's context and humanitarian challenges (15%)
- Value for money (15%)
8. Working Arrangements
- The consultant will work closely with the Global AI team and the data engineering/DevOps specialist
- Primarily remote work
- Regular coordination with RRM consortium partners
- All code and documentation will be stored in the project's version control system
- The consultant will report to the AI Lead
9. Intellectual Property
All models, code, and documentation developed during this consultancy will become the property of NRC and will be made available to consortium partners. Selected components may be made available under appropriate open-source licenses as determined by the project leadership.
10. Application Deadline
Applications must be submitted to [ sd.procurement@nrc.no ] by [11th of May 2025].
Interview dates:
Bids must include the following:
· The proposal includes outline of the evaluation framework and methods, including comments on the TOR, proposed time frame, and work plan (bids over 3 pages will be automatically excluded).
· Proposed evaluation budget
· CVs
